Introduction

The ongoing SARS-CoV-2 infection pandemic had been associated with huge morbidity, mortality and economic burden. By 23 October 2022, over 624 million people have had documented COVID-19 infection worldwide, and more than 6.5 million died [1]. The level of the pandemic impact on almost every aspect of human life was extremely high necessitating extended lockdowns, strict precaution measures and restrictions of local and international transportations. These unprecedented effects were of such magnitude as to prompt extraordinary attempts at producing potent vaccines at a speed that is now considered a historical record. After the introduction of the vaccines, overwhelming data from different studies indicated that vaccines were capable of inducing high rates of humoral and cellular responsiveness besides providing clinical protection against the infection, hospitalizations and mortality [2, 3].

Subsequently, reports from multiple large-scale researches indicated that certain people, most notably the elderly, the immunocompromised and those with certain comorbidities (e.g. hypertension, lung disorders, etc.) are significantly at higher risk of infection and specially its sequelae [4]. One of these at risk populations are chronic kidney disease patients with reports indicating high rates of hospitalizations and deaths due to COVID-19 infection. On the other hand, controversial reports have been published on the efficacy of COVID-19 vaccinations in this and other subgroups of patients with immunocompromised states indicating suboptimal responses to standard vaccination in either laboratorial assays or clinical settings [4]. Nevertheless, subsequent introduction of booster vaccinations has substantially changed the current knowledge with reports of higher protection levels offered by booster vaccination [5]. In this study, the aim had been to investigate the effects of booster vaccination on the humoral and cellular immune response to SARS-CoV-2 spike antigen in patients under renal replacement therapy (RRT) including patients undergoing conventional dialysis (hemodialysis (HD) or peritoneal dialysis (PD)) and kidney transplant recipients (KTR).

Methodology

Literature search strategy

A systematic review of the literature has been conducted on August 2022 in the Medline/Pubmed and Europe PMC search engines in quest of studies reporting their experience with the efficacy of booster vaccination against SARS-CoV2 infection in patients under RRT. The literature search had been lastly updated on September 26, 2022 to include the potential new publications or finding the potential missing ones. The found studies have been reinvestigated for their citations through Google Scholar searches to find more studies potentially missed by the original method. Search terms mainly included “Third dose + COVID-19 vaccination + dialysis”, “Booster + COVID-19 vaccine + kidney transplant”, and different combination of terms including word-level substitutions of terminologies (e.g. SARS-CoV-2 for COVID-19, renal for kidney, hemodialysis, “peritoneal dialysis” or “renal replacement therapy” for dialysis, “third dose” or “forth dose” for booster, mRNA-1273 for Moderna, and so forth). Due to the subspecial target population to the systematic review, no restriction was set for inclusion regarding the studies’ publication dates, study design (prospective or retrospective), language (to the extent they could be comprehended by the author—finally all the reviewed reports were in English), inclusion of a control group (or not), vaccine type/formula, vaccine brands, laboratory assays employed, booster order (third, fourth, etc.), status regarding peer-reviewing (pre-print studies have also been included) was applied. Sample size had a restriction criterion: to be included, studies should have had at least ten reported cases. As well, in studies reporting heterogeneous patient populations (e.g. solid organ transplants), data were included into the analysis only if the regarded information for the study target populations (HD, PD or KTR) could be discretely retrieved.

Study selection

The literature screening was initially conducted reviewing the article titles. Studies with relevant titles were further reviewed by their abstracts, and in case the abstracts implicated data of interests, the fulltexts were reviewed. Finally, meta-analysis of data were performed from the studies meeting the following criteria: human participants with end-stage renal disease under RRT methods (HD, PD, KTR) receiving standard SARS-CoV-2 vaccination who had received at least one booster dose from any type or any manufacturer. Studies containing data from RRT patients were included into the meta-analyses irrespective of the inclusion of healthy control groups.

Data extraction

Supplementary tables1 and 2 summarize the data of interest that have been sought and collected for each study on the KTR and dialysis patients, respectively. Data were extracted according to a designed checklist in Microsoft Excel. The key information were re-stratified according to the checklist, and reviewed twice by the author. The extracted data included the RRT setting (KTR, HD/PD), study design (retrospective or prospective), country of study, vaccine type (i.e. mRNA, viral vector and whole inactivated virus) and brands (i.e. BNT162b2 (Pfizer), mRNA-1273 (Moderna), ChAdOx1, Ad26.COV2.S (J&J), and CoronaVac), study population, immunosuppressant employed, predefined threshold levels of antibody response (cutoff for seroconversion), poor response to the standard vaccination (including non-responders (no seroconversion) and suboptimal responders); data for each extracted separately as well), time of antibody testing after booster dose, assays employed for the antibody testing, T cell reaction, seroconversion before and after the booster dose (one study employed data from a control group in spite of pre-booster dose measure [12]), any rejection episodes or graft failure after booster vaccination, baseline serum creatinine levels for responders and non-responders, and breakthrough infections after booster dose vaccination. Wherever possible, only naïve patients were included into the analyses and those with a history of infection were excluded. Studies that had only included non- or suboptimal-responders to the standard two-dose vaccination were excluded from the epidemiological reports of booster dose immunogenicity and were included into separate meta-analyses of patients with non- and suboptimal-response to the conventional two-dose schedules. Moreover, data from patients under highly potent monoclonal antibodies especially belatacept, were not included into the main epidemiological analyses wherever possible, but were secured for inclusion into the meta-analyses of the immunosuppressant effects.

Outcomes’ definition

Primary outcomes

The main primary outcomes of interest in this study were the differential rates of breakthrough infections, hospitalizations and death due to COVID-19 infection after standard- versus booster dose vaccinations in patients on RRT. Although individual studies had occasional data on some of the above-mentioned factors, none of which met the criteria required for inclusion into meta-analyses. The only data existing at an enough volume for meta-analysis was the breakthrough infection rates just after booster vaccination. Besides, seropositivity rates post-booster vaccination were analyzed for the renal patients compared to that in healthy controls. Seroconversion rates early after a booster dose (about 4 weeks after the third or fourth dose) was the next primary outcome index. For this purpose, serum antibody levels or reports of seropositivity status just before the booster doses and one month after had been collected for the efficacy analysis. Whenever data were unavailable on the seropositivity before the booster dose, the epidemiology of post-booster seropositivity were pooled and meta-analyzed separately. As well, data of seroconversion 1 month after the standard (2-dose) vaccination were collected for comparing with the post-booster data to evaluate the potential cumulative effect of the booster doses on inducing seroconversion (versus only compensating for the waning immunity). Another measure of primary outcome was T cell-specific immunity rates in response to booster dose vaccination against SARS-CoV-2.

Secondary outcomes

In the secondary outcomes, factors affecting the response to booster vaccination were sought. The differential impact of booster vaccination in KTR patients versus those under dialysis, the disparity in response to the booster dose regarding the type (mRNA versus vector vaccines) and brands (BNT162b2 vs. mRNA-1273) of the administered vaccines in different contexts (regarding the study participants’ response to the standard vaccinations) were some of the predefined secondary outcomes. Finally, the differential impact of the immunosuppressant types, which were employed for preventing kidney allograft rejection, on the seroconversion rates was also investigated.

Data synthesis and statistical analysis

DerSimonian and Laird random effects model was used to estimate the pooled Odds ratios and corresponding 95% confidence intervals for the primary and secondary outcomes of interest. An Odds ratio > 1 indicates that the RRT participants had a higher rate of achieving response (i.e. seroconversion or T cell response) post-booster vaccination compared to the control (pre-booster, post-dose 2 vaccine or healthy controls). Only in one case (RRT patients versus healthy controls), meta-analyses of randomized controlled trials could be achieved, and in the other subgroup analyses, the comparisons were made in the same subjects between measurements at different timepoints (i.e. 1 month after booster COVID-19 vaccination compared with own serostatus before booster vaccination at two timepoints: right prior to the booster dose or one month post-dose 2 of the standard protocol). In one study, the comparisons were between post-booster cases and controls’ measures at the same timepoint [12]. Statistical heterogeneity of the meta-analyses was assessed using χ2 test and I2 statistic, with p values < 0.10, or the I2 statistic ≥ 50% as definitions for significant heterogeneity. To avoid biases, studies containing significant bias in their included populations or approaches were excluded in each subsection meta-analysis, regarding the analysis purpose. For example, in the overall response to the booster doses, studies including only patients with impaired response to the standard vaccination protocol or those including only patients under highly potent monoclonal immunosuppression were excluded. That same criteria were applied to other subsection analyses including the vaccine type or brand, and immunosuppression effect meta-analyses.

Subgroup analyses were conducted to determine whether the results were influenced by the type of RRT, vaccine types or brands, and the levels of response to the standard vaccination (i.e. normal, weak- or non-responders). Stratifications were then used to estimate effect sizes in different subgroups, and to compare the differences between the respective estimates. Since the serum anti-SARS-CoV-2 antibody levels had been measured through different methods and assays, they were considered to be amenable to statistical pooling for meta-analysis.

In subgroup analyses where there was no pre-booster results for comparisons (e.g. breakthrough infections, or overall post-booster seroconversion rates irrespective of the existence of pre-booster measurements), the absolute risk was measured as a proportion from 0 to 100%. All statistical analyses were conducted using STATA vs.17 (StataCorp.) and p value of < 0.05 were considered statistically significant.

Results

Figure 1 illustrates flowchart of the study selection and review process. Overall 58 studies [6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36,37,38,39,40,41,42,43,44,45,46,47,48,49,50,51,52,53,54,55,56,57,58,59,60,61,62,63] with a total number of 8596 participants were finally selected and included into the systematic review and meta-analysis. The included trials had been conducted in Austria, Brazil, Canada, Denmark, France, Germany, Israel, Italy, Japan, Norway, Poland, Spain, Thailand, and USA. Characteristics of the included trials are summarized in supplementary tables 1 and 2. No rejection episodes or graft failure after booster dose was reported by any of the reviewed studies.

Fig. 1
figure 1

Flowchart

Full size image

Primary outcome

Figure 2 illustrates the forest plot of the meta-analysis of the seroconversion rates in response to booster vaccination in renal patients versus healthy controls. Compared to the healthy controls, the KTR patients represented significantly lower rates of seroconversion for SARS-CoV-2 post-booster vaccination (Odds ratio (95% CI), 0.05 (0.01–0.28; I2 = 0%), while the associated rates were not significant for patients under dialysis (0.68 (0.06–7.13; I2 = 0%); the between group difference (dialysis vs KTRs) did not reach significance level (p = 0.08).

Fig. 2
figure 2

Meta-analysis of seropositivity rates post-booster vaccination in RRT patients compared to healthy controls

Full size image

As shown in Fig. 3, compared to the pre-booster measurements, seropositivity was significantly augmented 1 month after booster vaccination for both the RRT subgroups (Odds ratio (95% CI) of 3.99 (3.16–5.04; I2 = 55.8%) and 6.53 (2.32–18.39; I2 = 69.6%) for the KTR and dialysis patients, respectively), with no significant difference between the two groups (p = 0.28). However, the overall post-booster dose seropositivity rate (irrespective of the pre-booster status) was significantly higher in the dialysis population than the KTR patients (0.88 (0.84–0.93; I2 = 86.4%) vs. 0.69 (0.64–0.73; I2 = 83.6%), respectively; p < 0.01) (Fig. S2). On the other hand, compared to the rates 1 month after the second dose, third dose vaccination was associated with significantly higher seropositivity at 1 month only in the KTR subgroup, with no significant difference detected for the dialysis patients (odds ratio (95% CI), 3 (1.82–4.95; I2 = 74.7%) and 1.42 (0.66–3.04; I2 = 0%), respectively) (Fig. S3).

Fig. 3
figure 3

Meta-analysis of seropositivity rates post-booster vaccination versus pre-booster measurements in RRT patients

Full size image

Due to data scarcity on reports of breakthrough infections post-standard (two-dose) vaccination, only reports of breakthrough infections post-booster vaccination were included into meta-analysis. Using the DerSimonian–Laird method, the overall estimated incidence [95% confidence interval (CI)] of breakthrough infection with SARS-CoV-2 infection was 0.06 (95% CI 0.04–0.08; I2 = 93.2%). As illustrated in Fig. S4, compared to the KTR patients, the respective estimates for patients under dialysis had a higher trend of breakthrough infections [0.05 (95% CI 0.02–0.08; I2 = 92.3% vs. 0.10 (95% CI 0.05–0.16; I2 = 70.5%); though significance levels was not achieved, p = 0.09)]. T cell-specific response rates post-booster vaccination were very diverse in different trials with an overall 1.65 (0.98–2.77) and 2.89 (1.47–5.72) times larger rates compared to pre-booster status for the KTR and dialysis patients, respectively (p = 0.2, figure S5). Since some of the studies had no reports of cellular response prior to the third dose, the overall rate of T-cell specific reactivity post-booster vaccination were pooled and meta-analyzed which returned an overall rate (95% CI) of 0.53 (0.31–0.76) with no significant difference between the KTR and dialysis patients (p = 0.24, Fig. S6). Other predefined primary outcome indices could not be achieved due to the lack of reported data.

Secondary outcome

Although there are large body of evidence on the magnitude of the effects imposed by different vaccine types or brands on the seroconversion rates, the issue in RRT patients has not been broadly discussed in the literature. Figure 4 illustrates the meta-analysis of response to booster dose vaccination compared to the pre-booster serostatus stratified by the vaccine type, indicating that both vector and mRNA vaccine types when administered as the booster could significantly augment seropositivity with no significant difference, while both are superior to the inactivated virus vaccine. Similar findings were reached when vaccine brands of the booster dose were investigated, representing no significant difference between different vaccine brands (notably between BNT162b2 (Pfizer) & mRNA-1273 (Moderna) for mRNA vaccines) (Fig. S7). Results were reproduced when the meta-analysis included stratification by the RRT types (i.e. KTR patients taking BNT162b2 vs. mRNA-1273, etc.; Fig. S8).

Fig. 4
figure 4

Meta-analysis of vaccine type effects on the seroconversion rates after booster vaccination against SARS-CoV-2 infection

Full size image

Poor response to the standard vaccination and its effect on the immunogenicity of the booster dose was the next subject of investigation. As illustrated in figure S9, primarily poor responders to the standard vaccination seroconverted at a rate (95% CI) of 0.77 (0.65–0.89; I2 = 77.9%) and 0.41 (0.36–0.47; I2 = 87%) for dialysis and KTR patients, respectively, with dialysis patients representing a significantly higher seroconversion rates than KTRs (p < 0.01). The significance of the vaccine types as the booster dose on seroconversion in this subpopulation was also investigated; mRNA and vector-based vaccine types were equally efficacious to induce seroconversion (0.51 (0.49–0.54; I2 = 0%) and 0.52 (0.23–0.82; I2 = 86.6%), respectively) while both were significantly more efficacious than inactivated virus vaccine (0.20 (0.16–0.23; I2 = 0%) (Fig. S10). Reanalysis after stratifying the data based upon vaccine brands did retuned no significant effect (Fig. S11).

Then poor-responders’ data were substratified again to investigated booster dose effects, separately for non-responders and suboptimal responders of the standard 2-doses. Dialysis patients in either group were significantly more likely to seroconvert in response to boosters compared to the KTRs (0.66 (0.58–0.73; I2 = 0%) vs. 0.42 (0.36–0.48; I2 = 52.1%), p < 0.01 Fig. S12, and 0.88 (0.83–0.94; I2 = 0%) and 0.33 (0.31–0.34; I2 = 0%), p < 0.01 Fig. S13, respectively).

Demographic data

No significant gender bias was found regarding the response to boosting vaccination (Fig. S1). Attempts on pooling data of other demographic factors including the baseline serum creatinine/eGFR levels, age, vintage on dialysis (or transplantation), and BMI failed because of either the inconsistencies in the measurement units (i.e. mg/dL versus µM for creatinine and different time units (i.e. years/months/days) for vintage statistics) or different statistical indices for reports [i.e. mean (SD) or median (IQR/range)]. Yet attempts have been made to provide an overview. In case of baseline serum creatinine levels, except for one study [17], all other reports indicated trends of lower serum creatinine values in seroresponders than in non-responders [53, 58, 61], with some of them reaching significance [18, 21, 60]. Vintage on RRT, however, tended to be longer in responders in all [17, 21, 32, 53, 61], but one [13] study. Likewise, in only one study [32], the BMI was higher in the non-responders, and in the remaining reports, seroconverters characterized with higher BMI trends [13, 17, 21, 61]. Age effect on the seroresponse was very inconsistent in different studies. While in three studies responders were inclined towards older ages [17, 32, 61], in three others, advanced age was associated with non-responsiveness [13, 21, 53].

Immunosuppressant effect

Figures S14-S33 summarize meta-analyses investigating differential effects of the immunosuppressant drugs on their serological response to vaccination. As is evident in the figures, the lowest booster vaccination response belonged to patients under belatacept, significantly lower than that under mycophenolate mofetil (MMF), tacrolimus, calcineurin inhibitors (CNi, i.e. tacrolimus and cyclosporine), Mammalian target of rapamycin inhibitors (mTORi, i.e. sirolimus and everolimus), and steroids, while the best response was detected for patients under treatment with mTORi significantly higher than that for MMF, tacrolimus or CNi, steroids, basiliximab, and belatacept. No other significant difference was detected regarding the type of immunosuppressant.

Discussion

This systematic review of 58 studies showed significant cellular and humoral immune response in patients under either type of RRT (i.e. dialysis, kidney transplants) after booster dose vaccination against SARS-CoV-2 infection. Although cellular or serological response does not essentially mean protection against infection, however, due to overwhelming epidemiological data on the prevalence of infection in patients with or without immune reactivity, findings of this study put high priority on boosting vaccination for these patients.

The detected rise in the seropositivity rates after boosting vaccination can be interpreted in two ways: compensating for the waning immunity after the standard two-dose vaccination, and/or through intensification of the response to the standard protocol. The literature suggests lower seroconversion rates after vaccination against SARS-CoV-2 in both dialysis and KTR patients than that for the healthy controls (27%, 88% and 100%, respectively), with a highly significant difference between the RRT methods [64]. In the current study, the post-booster seropositivity rates were 41%, 77% and 100% for the KTR, dialysis and healthy controls with patients in both RRT subgroups significantly augmenting seroconversion rates compared to pre-boosting measurements, yet it was only the KTR and not dialysis patients whose seroconversion rates went significantly beyond the levels measured after the standard two doses (Fig. S3). That means, booster vaccination in the dialysis patients well compensates for the waning immunity after standard dose vaccination, while in the KTRs, it actually surpasses counteracting the waning immunity and significantly intensifies the seroconversion provided by the standard protocol at the first place. This warrants consideration of the third dose for all the KTR patients as part of their standard vaccination protocol, while more doses (i.e. fourth or more) could be considered as boosters.

Indeed, the most important measures in the evaluation of the vaccine efficacy against SARS-CoV-2 infection are hospitalizations and deaths. However, due to data scarcity on these indices in the reviewed reports, the next important milestone achievable through this systematic review was breakthrough infection. Although compared to the KTRs, patients under dialysis therapy represented significantly superior humoral response to vaccine boosting, and also they were the only RRT subgroup that experienced significant augmentation of T cell response post-booster dose, the incidence of breakthrough infection in these patients was almost twice as high as that of the KTRs. This might undermine the general concept that humoral or cellular response can be interpreted to clinical benefits, however, a plausible explanation for this seemingly controversy is the disparity in the exposure, with dialysis (especially hemodialysis) patients being at higher risk of exposure to the virus, due to their compelling frequent attendance to the dialysis facilities. Moreover, the epidemiology of breakthrough infections should be interpreted with caution, because the screening methodologies were not consistent among different studies, with disparities in the detection methods as well as their approaches with some of them systematically testing the patients for asymptomatic infections, while some others only sought to test and report the symptomatic individuals. Follow-up duration was also significantly different between the studies ranging from less than 2 to over 8 months post-booster vaccination [7, 51].

The SARS-CoV-2 vaccine constructs have also been proposed to play major roles in the seroconversion rates, with the mRNA vaccines significantly eliciting higher immunogenicity than the other types of vaccines, most notably the viral vectors [65], although there are reports indicating heterogeneous mRNA-vector vaccination could be as immunogenic as homologous mRNA vaccination [66]. The current study on patients under RRT is consistent with the latter, where homologous or heterologous booster dose vaccination with the vector or mRNA vaccines were equally effective in inducing seroconversion in RRT patients, whereas inactivated virus vaccine was less effective. Regarding the brands of the vaccines, the literature indicates superiority in effectiveness for mRNA-1273 versus BNT162b2 [67]. Although due to the shortage of evidence, the current systematic review could only investigate the seroconversion rates after booster vaccination, no vaccine brand predilection was found in this regard.

Besides the overall responsiveness to booster dose vaccination in immunocompromised patients, the idea of how the non-(or weak-) responders to the standard vaccination would respond to boosting is a distinct and critical idea. As had been observed for the overall analyses, while both the RRT subgroups with poor-response to the standard vaccine protocol significantly augmented seroconversion, the dialysis patients were more likely to respond to the booster vaccination than the KTRs (Fig. S9), and this disparity remained significant even after excluding the two studies that had employed inactivated virus vaccines for their KTR population [42, 47] (data not shown). As had been observed in baseline analyses, the efficacy of booster vaccination in the poor responders was equivalent with respect to the vaccine types or brands, except for the inactivated virus vaccine which represented relatively less effectiveness (Fig. S8).

The immunosuppressive agents have also been shown to significantly impact response to vaccination [68]. In the current study, most of the immunosuppressant drugs represented no different immunogenicity in response to booster vaccination in the KTRs with two exceptions: compared to all the other drugs, mTORi agents offered the best vaccine response while using belatacept was associated with the least seroconversion rates. This finding has clinical implication with giving priority to specific immunosuppressant drugs to be employed or avoided at the time of vaccination.

The current study is associated with a lot of limitations. Most notably, the review was not on the randomized controlled trials. Although a number of the studies had used healthy controls, but the main body of the review included observational trials in which the seropositivity rates were investigated at pre- versus post-booster dose vaccination timepoint. There might be more confounding factors regarding the outcomes of interest. For example, not essentially every patient who received the primary vaccination also received boosters, and there might be some selection biases towards boosting vaccination in the less responsive and/or more vulnerable subjects. That same limitation might apply to the differential effects of vaccine types and brands: There might be skewed selection in favor of the particular vaccine types/brands that have been previously shown to be more reactogenic in the general population, to be employed in the most vulnerable and immunocompromised patients. But since the majority of the studies were consistently using one vaccine brand in all their patients, this limitation is not likely to significantly confound the results. There was also different time intervals between the standard and booster doses in different studies which could confound the comparisons. On the other hand, different studies were using different assays for antibody detection, and the antibody thresholds set to determine seropositivity was also inconsistent between the different studies, although it is noteworthy that the same outcome indices had been essentially used for both the measurements (i.e. pre- and post-boosting) in each study. Limitations are also applicable to the meta-analyses of the immunosuppressant effects on the seroconversions: besides the inconsistencies in dosing schedules, the investigated immunosuppressant drugs were not used in isolation and mostly were used in combination to other drugs. So confounding effects of the other drugs should also be considered while interpreting the results.

In conclusion, this study signifies the safety and utmost importance of the booster vaccination in patients under renal replacement therapy and suggests a third vaccine dose to be included as the standard vaccination protocol to the kidney transplant patients. As well, since dialysis patients represented higher breakthrough infections compared to the KTRs despite higher seroconversion rates, the same idea (standardization of booster dose vaccination) is also reasonable for application in this patient population. Except for the inactivated virus vaccines, no priority was found regarding the vaccine types or brands to be used as booster doses in this patient population. Future studies targeting hard outcomes, including the SARS-CoV-2-associated hospitalizations and mortality are recommended.