Humoral and Cellular Response of Transplant Recipients to a Third Dose of mRNA SARS-CoV-2 Vaccine: A Systematic Review and Meta-analysis

Background.

High rates of nonresponse to 2 doses of mRNA severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) vaccine have been reported in transplant recipients. Several studies have investigated the efficacy of a third dose in this population. However, efficacy remains unclear, as response rates vary across studies. Therefore, we conducted a systematic review and meta-analysis to determine the efficacy of a third dose of any mRNA SARS-CoV-2 vaccine in transplant recipients.

Methods.

Preferred Reporting Items for Systematic Review and Meta-Analysis reporting guidelines (PROSPERO:CRD42021281498) were followed. Medline, Embase, and CENTRAL were searched from inception to December 2, 2021, without restrictions. All full-text studies reporting on the efficacy of a third dose of any mRNA SARS-CoV-2 vaccine in pediatric and adult transplant recipients were included. The National Institutes of Health quality assessment tool for case series and the Cochrane risk of bias tool determined study quality. Meta-analysis was performed via the DerSimonian-Laird random-effect model.

Results.

Of 84 records, 12 studies totaling 1257 patients met inclusion criteria. One study was a randomized controlled trial, whereas all other studies were observational. Across 7 studies (801 patients), humoral response after 3 doses was observed in 66.1% (95% confidence interval, 62.8%-69.4%; I² = 0%) of transplant recipients. Triple immunosuppression, mycophenolate, antiproliferatives, and belatacept use were associated with reduced odds of humoral response in studies reporting multivariate analyses. Transplant recipients receiving a third dose displayed higher levels of neutralizing antibodies to SARS-CoV-2 variants (Alpha, Beta, and Delta) compared with placebo.

Conclusions.

A third dose SARS-CoV-2 mRNA vaccine should be strongly considered in transplant recipients. Limitations included lack of controlled studies and clinically relevant thresholds to determine response to vaccination.

INTRODUCTION

Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), responsible for coronavirus disease 2019 (COVID-19), is a zoonotic viral pathogen that has become a global health concern. Low seroconversion rates (54%) after the second dose of mRNA SARS-CoV-2 vaccine have been reported in transplant recipients,¹ and severe cases of COVID-19 have been described in transplant recipients who have received 2 doses of vaccine.² Given the high mortality rate of COVID-19 in transplant recipients,^3,4 health authorities have issued recommendations to administer a third dose in this population.¹ However, the efficacy of a third dose of mRNA SARS-CoV-2 vaccine in transplant recipients varies across studies,^5-7 with some studies reporting humoral response after 3 doses as low as 6.4%.⁵ Additionally, the relative efficacy of a third dose according to transplant type, mRNA vaccine, and factors associated with immune response remains unknown. The literature is also highly reliant on humoral response and lacks other indicators that are more suggestive of immunity, specifically cellular response and neutralizing antibody assays.⁸ As neutralization level is highly predictive of immune protection⁹ and T cells reduce viral loads and disease even when neutralizing antibody levels are low,¹⁰ these outcomes are of vital importance. Finally, these vaccines have demonstrated decreased efficacy against variants of concern (VOC),^11,12 which are the dominant circulating strains in the community,¹³ and data on vaccine efficacy with respect to VOC in transplant populations remain unknown. Therefore, a systematic review and meta-analysis is required to determine the humoral and cellular response of transplant recipients to 3 doses of mRNA SARS-CoV-2 vaccine, according to viral strain, and factors associated with immunity in this population.

MATERIALS AND METHODS

This review was reported in accordance with the Preferred Reporting Items for Systematic Review and Meta-Analysis¹⁴ (Tables S1 and S2, SDC, http://links.lww.com/TP/C610). Our protocol was registered on PROSPERO (CRD42021281498; September 24, 2021).

Search Strategy and Study Selection

Medline, Embase, and Cochrane Central Register of Controlled Trials (CENTRAL) were searched from inception to December 2, 2021, without language or geographic restrictions, using a search strategy developed by a librarian (R.S.) specializing in systematic searches (Table S3, SDC, http://links.lww.com/TP/C610) according to the Peer Review of Electronic Search Strategies guidelines.¹⁵

Two independent reviewers (A.J.M.B. and H.B.M.) screened all records in Rayyan software,¹⁶ with supervising author providing consensus (D.S.A.). All studies reporting on the efficacy of a third dose of any mRNA SARS-CoV-2 vaccine in pediatric and adult transplant recipients were included. Studies reporting on nontransplant recipients (eg, Car-T-cell recipients) or studies reporting other doses (ie, first or second doses) were excluded. Studies were not required to be comparative. Randomized trials, case-control studies, cohort studies, cross-sectional studies, and single-arm studies were included, whereas commentaries, editorials, conference abstracts, reviews, gray literature, preprints, and guidelines were excluded. Manual searches through reference lists of included articles were conducted to capture any other relevant literature not captured by our search strategy.

Outcomes

Our primary outcomes consisted of the humoral and cellular response of transplant recipients of any age and transplant type to a third dose of any mRNA SARS-CoV-2 vaccine. Humoral response was considered positive if antibody titers exceeded the study’s threshold. Cellular response was considered positive if the ratio of SARS-CoV-2–specific T cells to T cells, respectively, exceeded the study threshold. If a study threshold was not provided, a threshold from another study using the same T-cell population was applied. Secondary outcomes included factors associated with immunity, adverse events, and safety.

Data Extraction

Two reviewers (A.J.M.B. and H.B.M.) independently extracted data using data extraction forms for all primary and secondary outcomes, with supervising author providing consensus (D.S.A.). Extracted data included study characteristics (authors, year, study design, patient age and sex, transplant type, time since transplant, vaccine type received, time since second dose, time of assessment of outcomes), details regarding humoral and cellular assays (antibody assay, measured antibody, threshold, cellular type, markers analyzed), and outcomes (prevalence of humoral and cellular response after 3 doses, prevalence of humoral and cellular response after 3 doses in patients without response to 2 doses, acute rejection, factors associated with humoral response, side effects, adverse events, and acute rejection episodes).

Risk of Bias Assessment

Two reviewers (A.J.M.B. and H.B.M.) independently assessed the risk of bias for all primary and secondary outcomes, with supervising author providing consensus (D.S.A.). For the synthesis of noncomparative data, such as humoral response in a single group, the National Institutes of Health quality assessment tool for case series was used (https://www.nhlbi.nih.gov/health-topics/study-quality-assessment-tools, accessed October 2, 2021). For comparative data, version 2.0 of the Cochrane Risk of Bias Assessment Tool for randomized trials was applied (low risk, some concerns, or high risk of bias).¹⁷ All poor-quality studies or studies with a high risk of bias were excluded from the qualitative or quantitative synthesis. Sensitivity analysis was performed according to study quality. Funnel plots were used to assess for publication bias.

Certainty of the Evidence

The Grading of Recommendations Assessment, Development, and Evaluation (GRADE)¹⁸ was used to assess the certainty of the evidence for all outcomes. Although there is no formal guidance for GRADE in systematic reviews of prevalence,¹⁹ a recommended framework was applied.²⁰ Two reviewers (A.J.M.B. and W.C.) independently assessed certainty of the evidence, with supervising author providing consensus (D.S.A.).

Statistical Analysis

Meta-analysis of the prevalence of humoral and cellular response of transplant recipients after 3 doses of mRNA SARS-CoV-2 vaccine was performed in OpenMetaAnalyst²¹ using the DerSimonian-Laird random-effect model.²² I² >50% represented substantial heterogeneity,²³ and heterogeneity was evaluated using the Cochrane Q-statistics with a significance level of P < 0.10.²⁴ P <0.05 was considered as significant for all other analyses. Prespecified subgroup analyses were performed according to transplant (ie, heart, lung, kidney) and vaccine type (ie, mRNA-1273 versus BNT162b2). Subgroup analyses were also performed according to study threshold and according to correlates of protection according to SARS-CoV-2 variant (wild type, alpha, delta).²⁵ The presence of publication bias was only assessed if >10 studies were available for each outcome.²⁴ Solid organ transplant recipients were analyzed separately from hematopoietic stem cell transplant (HSCT) recipients.

For analysis of associations, the odds ratios (ORs) of predictive factors for humoral response could not be pooled as the multivariable logistic regression performed across studies accounted/adjusted for different covariates. Thus, associations were synthesized qualitatively.

All data are presented with 95% confidence intervals (CIs). All tests were 2-sided and statistical significance was based on the 95% CIs excluding the null.

RESULTS

Study Selection

Of 84 records, 12 studies^5-7,25-33 were included encompassing 1257 patients (Figure S1, SDC, http://links.lww.com/TP/C610).³⁴ Along with editorials and commentaries, 3 studies were excluded in the full-text phase. One study reported prevalence of humoral response in a single cohort of transplant recipients receiving 2 or 3 doses,³⁵ another study did not adequately characterize its study populations (ie, various data such as time since transplant and immunosuppressive regimens were not reported) and did not meet threshold for quality,³⁶ and another study applied inconsistent thresholds in their data reporting.³⁷

Study Characteristics

Study and patient characteristics are summarized in Table 1. All patients were aged >18 y, and the range of medians/means was 54.8 to 66.6 y. In most (8/10) studies, the third dose was administered 2 mo after the second dose (range of medians/means, 21–168 d), with outcome measurement occurring 1 mo afterward (range of medians/means, 14–52 d). All studies used defined thresholds to determine humoral and cellular response, although assays and thresholds varied across studies (Table 2). Transplant types included 878 kidney (69.8%), 147 heart (11.7%), 101 liver (8.0%), 42 allogenic HSCT (aHSCT; 3.3%), 30 lung (2.4%), and 59 other transplants (4.7%), such as pancreas and combined transplants.

TABLE 1.

Characteristics of included studies

Author (location)	Study design	Age, y (%M)	Transplant	Time since transplant (y)	Vaccine	Time since second dose	Time of assessment	Humoral assay (threshold)
Hall et al7(Canada) and Kumar et al27 (Canada)	RCT	66.6 [63.3–71.4] (74%)	29 Kidney29 Lung20 Liver18 Heart24 Other	3.16 [1.71–6.12]	mRNA-1273	60 d	30 d	Antispike-RBD IgG(≥100 U/mL)
Kamar et al28(France)	Single arm	57.6 ± 17.2 (69.7%)	76 Kidney12 Liver7 Other	8.1 ± 1.0	BNT162b2	61 ± 1 d	30 d	Anti–SARS-CoV-2(S/Co>1.1)
Del Bello et al29(France)	Single arm	54.8 ± 12.8 (65.4%)	277 Kidney69 Liver33 Heart17 Other	6.8 ± 8.1	BNT162b2	59 d [47–67]	30 d	Anti–SARS-CoV-2(S/Co>1.1)
Chavarot et al5 (France)	Single arm	63.5 [51–72] (58%)	62 Kidney	4.0 [2.1–6.6]	BNT162b2	69.5 d [40–84]	28 [28–33]	SARS-CoV-2 spike protein(≥50 AU/mL)
Peled et al30(Israel)	Single arm	61.0 [49.8–68.0] (71%)	96 Heart	6.3 [3.5–13.6]	BNT162b2	168 ± 18 d	18 d	Antispike-RBD IgG(>12.6 geometric titer)
Westhoff et al31 (Germany)	Single arm	59.5 ± 11.6 (80%)	10 Kidney	5.2 ± 4.7	mRNA-1273	64 ± 15 d	14 d	SARS-CoV-2 spike and nucleocapsid protein(≥0.75 U/mL)
Benotmane et al6(France)	Single arm	57.6 [49.6–66.1] (61.6%)	159 Kidney	5.3 [1.9–11.1]	mRNA-1273	51 d [48–59]	28 d [28–32]	Antispike-RBD IgG(>50 AU/mL)
Redjoul et al32 (France)	Single arm	59 [50–64] (65%)	42 Allogenic HSCT	<1 y (22)>1 y (20)	BNT162b2	51 ± 22 d	26 ± 6 d	Antispike-RBD IgG(≥4160 AU/mL)
Masset et al33(France)	Single arm	63.7 ± 11.7 (63.2%)	124 Kidney12 Other	9.5 ± 8.1	BNT162b2	50 d ± NR	28 d [27–33]	Antispike-RBD IgG(>250 UI/L)
Bertrand et al34 (France)	Single arm	63.6 ± 15.7 (60%)	80 Kidney	7.3 [3.4–14.1]	BNT162b2	67.5 d[57–70]	At least 28 d	Antispike-RBD IgG(>50 AU/mL)
Massa et al26 (France)	Single arm	58 [47.1–66.1] (72.1%)	61 Kidney	4.5 [1.8–11.3]	BNT162b2	21 d	28 d	Antispike-RBD IgG(>50 AU/mL)

Interquartile range is presented in square brackets.

AU, Arbitrary units; HSCT, hematopoietic stem cell transplantation; IgG, immunoglobulin G; IU, international units; %M, % male; NR, not reported; RBD, receptor-binding domain; RCT, randomized controlled trial; SARS-CoV-2, severe acute respiratory syndrome coronavirus 2; S/Co, signal to cutoff; U, units.

TABLE 2.

Detailed description of humoral and cellular assays across studies

Study	Cellular assay(s)	Threshold (s)	Humoral assay (s)	Antibody measuring	Threshold (s) positive
Del Bello et al29	Not reported	Not reported	Viral Ab titer: Wantai microplate ELISA	Anti–SARS-CoV-2 antibodies	S/Co>1.1
Hall et al7 andKumar et al27	Intracellularcytokine stainingMarkers analyzed: IFN-gamma/IL-2Cells analyzed: CD4+, CD8+ cells	<0.01(absolute values were reported)a	Viral Ab titer: anti–SARS-CoV-2 spike enzyme immunoassay (Roche Elecsys)	Antispike-RBD IgG	≥100 U/mL
			Neutralizing antibody test:(i) SARS-CoV-2 surrogate virus neutralization test(ii) SARS-CoV-2 pseudovirus neutralization assay (Kumar et al27 only)	(i) Antispike-RBD IgG(ii) N/A	(i) ≥30%(ii) Absence of 50% neutralization with undiluted serum
Kamar et al28	Not performed	Not reported	Viral Ab titer: Wantai microplate ELISA	Anti–SARS-Cov-2 antibodies	S/Co>1.1
Redjoul et al32	Not performed	Not reported	Viral Ab titer:Abbott Architect SARS-CoV-2 IgG Quant II assay (Abbott, Sligo, Ireland)	Antispike-RBD IgG	≥4160 AU/mL
Westhoff et al31	Intracellularcytokine stainingMarkers analyzed: IL-2/IL-4/IFN-γ/TNF-α/GrzBCell analyzed: CD4+CD154+CD127+and CD8+ CD137+	>0%	Viral Ab titer: measure using Elecsys Anti–SARS-CoV-2-S (Roche, Mannheim, Germany)	SARS-CoV-2 spike and nucleocapsid protein	≥0.75 U/mL
			Neutralizing antibody test: VSVG (FLuc) pseudovirus system bearing the SARS-CoV-2 spike protein	SARS-CoV-2 spike protein	≥0 AU
Benotmane et al6	Not reported	Not reported	Viral Ab titer: ARCHITECT IgG IIQuant test (Abbott)	Antispike-RBD IgG	>50 AU/mL
Chavarot et al5	Not reported	Not reported	Viral Ab titer: SARS-CoV-2 IgG II Quant antibody test (Abbott, USA)	SARS-CoV-2 spike protein	≥50 AU/mL
Masset et al33	Not performed	Not reported	Viral Ab titer:(i) Chemiluminescent microparticle immunoassay (Abbott Architect)(ii) Chemiluminescence immunoassay (Siemens Atellica)(iii) Electrochemiluminescence immunoassay (Roche Elecsys)	Antispike-RBD IgG	>250 IU/L
Peled et al30	T lymphocytes(T-cell markers are not described)	Not reported	Viral Ab titer: in-house ELISA	Antispike-RBD IgG	>12.6 geometric titer
			Neutralizing antibody test: GFP reporter-based VSVG backbone coated with the SARS-CoV-2 spike protein	SARS-CoV-2 spike protein	>1024 geometric titer
Bertrand et al34	Intracellularcytokine stainingMarkers analyzed: IFN-gammaCells analyzed: CD3+ T cells	S1 > 25 SFC/106 CD3+ T cellsS2 > 40 SFC/106CD3+ T cells	Viral Ab titer: ARCHITECT IgG II Quant test (Abbott)	Antispike-RBD IgG	>50 AU/mL
Massa et al26	ELISpot Path Human IFN-g (SARS-CoV-2, S1 scan+S2N+SNMO) ALP assay (MABTECH) measured IFN-gamma–secreting SARS-CoV-2–specific T cells per 106 PBMCs	None	Viral Ab titer: Abbott SARS-CoV-2 IgG assay	Antispike-RBD IgG	>50 AU/mL
			Neutralizing antibody test: V-PLEX! SARS-CoV-2 Panel 7 (ACE2) Kit (MSD) was used to measure binding of the SARS-CoV-2 full-length Spike protein to soluble ACE2 receptor	Anti–full-length SARS-CoV-2 spike protein antibodies	Noneb

^aThreshold from Westhoff et al³¹ was applied in meta-analysis.

^bThreshold from Kumar et al²⁷ was applied in meta-analysis.

Ab, antibody; ACE2, angiotensin-converting enzyme 2; ALP, Alkaline phosphatase; AU, Arbitrary units; ELISA, enzyme-linked immunosorbent assay; GFP, Green fluorescent protein; IFN, interferon; IgG, immunoglobulin G; IL, interleukin; IU, international units; MSD, Meso scale diagnostics; N/A, not available; PBMC, peripheral blood mononuclear cells; RBD, receptor-binding domain; S1, 15-mer peptide pools spanning the sequence of SARS-CoV-2 spike protein S spanning the N-terminal; S2, 15-mer peptide pools spanning the sequence of SARS-CoV-2 spike protein S spanning the C-terminal; SARS-CoV-2, severe acute respiratory syndrome coronavirus 2; SFC, spot-forming cells; S/Co, signal to cutoff; TNF, tumor necrosis factor; U, units; VSVG, vesicular stomatitis virus glycoprotein.

Humoral Response

Although all studies measured response against SARS-CoV-2 spike protein, assays differed across studies—depending on the manufacturer and the kit sensitivity, the threshold cutoff varied. When the prevalence of humoral response after 3 doses was pooled, high heterogeneity (I² = 97.6%) was observed (Figure S2, SDC, http://links.lww.com/TP/C610). This heterogeneity was found to be because of one study,⁵ which reported 6.4% response in a cohort of kidney transplant recipients (KTRs) receiving belatacept, and was removed as an outlier. Across 7 studies^{7,26,28-30,33,34} (801 patients), the prevalence of humoral response after 3 doses was 66.1% (95% CI, 62.8%-69.4%, I² = 0%; Figure 1A), with similar rates found across studies of kidney (61.7%; 95% CI, 53.7%-69.7%; I² = 0%)^26,34 and heart (66.7%; 95% CI, 57.2%-76.1%)³⁰ transplant recipients (test for subgroup differences, P = 0.56; Figure 1B). Prevalence of humoral response according to mRNA vaccine was higher in transplant recipients who received 3 doses of BNT162b2 (66.9%; 95% CI, 63.5%-70.3%; I² = 0%)^{26,28-30,33,34} versus 3 doses of mRNA-1273 (55.0%; 95% CI, 42.4%-67.6%; I² = N/A)⁷ (test for subgroup differences, P = 0.07; Figure 1C). Across 9 studies^{6,7,26,28-31,33,34} (747 patients), humoral response to a third dose after humoral nonresponse to 2 doses was observed in 45.7% of patients (95% CI, 41.8%-49.7%; I² = 13.1%; Figure 2A), with similar rates found across studies of kidney (43.2%; 95% CI, 33.6%-52.9%; I² = 45.5%)^6,26,31,34 and heart (54.3%; 95% CI, 42.6%-66.0%)³⁰ (test for subgroup differences, P = 0.30; Figure 2B). In aHSCT recipients, prevalence was 47.6% (95% CI, 32.5%-62.7%).³² In one randomized controlled trial (RCT),⁷ a greater number of patients in the mRNA-1273 group (55%) demonstrated humoral response compared with the placebo group (18%; risk ratio = 3.1; 95% CI, 1.7-5.8; P < 0.001). No differences in humoral response after humoral nonresponse to 2 doses were observed according to mRNA vaccine (mRNA-1273^6,7,31 49.6%; 95% CI, 43.0%-56.1%; I² = 0% versus BNT162b2^{26,28-30,32-34} 43.9%; 95% CI, 38.6%-49.2%; I² = 15.6%; test for subgroup differences, P = 0.19; Figure 2C). Notably, humoral response after 3 doses was observed in 52 of 64 patients (81.2%; 95% CI, 71.7%-90.8%) who had a weak humoral response and 26 of 95 patients (27.4%; 95% CI, 18.4%-36.3%) who had no humoral response to 2 doses.⁶ Prevalence of humoral response after 3 doses of any mRNA SARS-CoV-2 vaccine did not vary according to study threshold (test for subgroup differences, P = 0.43; Figure S3, SDC, http://links.lww.com/TP/C610). Prevalence of humoral response after 3 doses also did not vary according to correlates of protection for the wild type, alpha variant, and delta variant (test for subgroup difference P = 0.17; Figure S4, SDC, http://links.lww.com/TP/C610).

FIGURE 1.

Prevalence of humoral response after 3 doses of any mRNA SARS-CoV-2 vaccine in transplant recipients (A, top panel), also according to transplant type (B, middle panel), and mRNA vaccine (C, bottom panel). CI, confidence interval; SARS-CoV-2, severe acute respiratory syndrome coronavirus 2.

FIGURE 2.

Prevalence of humoral response after 3 doses of any mRNA SARS-CoV-2 vaccine in transplant recipients that did not display a humoral response to 2 doses of an mRNA SARS-CoV-2 vaccine (A, top panel), also according to transplant type (B, middle panel), and mRNA vaccine (C, bottom panel). CI, confidence interval; SARS-CoV-2, severe acute respiratory syndrome coronavirus 2.

Cellular Response

Five studies reported cellular response.^{7,26,30,31,34} Intracellular cytokine staining was used as a marker to measure cellular response, and markers analyzed within the studies largely varied. One study²⁶ was not able to be included in the meta-analysis because it reported the number of SARS-CoV-2–specific T cells per peripheral blood mononuclear cells. The percentage of patients achieving cellular response after 3 doses (75.3%; 95% CI, 66.6%-83.9%; I² = 24.5%)^7,30,31,34 (Figure 3A) was significantly higher (P = 0.0001) compared with after 2 doses (49.3%; 95% CI, 39.5%-59.1%; I² = 0%)^30,31,34 (Figure 3B). Among patients with cellular nonresponse to 2 doses, 57.8% (95% CI, 30.0%-85.6%; I² = 64.5%)^30,31,34 displayed cellular response after 3 doses (Figure 3C).^30,31,34 In one RCT,⁷ median SARS-CoV-2–specific T-cell counts after 3 doses of mRNA-1273 were greater than placebo (432 versus 67 cells per 10⁶ CD4⁺ T cells, 95% CI for the between-group difference, 46–986).

FIGURE 3.

Prevalence of cellular response after 3 doses of any mRNA SARS-CoV-2 vaccine in transplant recipients (A, top panel), after 2 doses of any mRNA SARS-CoV-2 vaccine in transplant recipients (B, middle panel), and after 3 doses of any mRNA SARS-CoV-2 vaccine in transplant recipients that did not display a cellular response to 2 doses (C, bottom panel). CI, confidence interval; SARS-CoV-2, severe acute respiratory syndrome coronavirus 2.

Neutralizing Antibody Response

Four studies reported neutralizing antibodies.^7,26,30,31 Although all studies measured response against SARS-CoV-2 spike protein, detected spike protein variants varied on the basis of the manufacturer. Furthermore, although 1 study²⁷ used a standard enzyme-linked immunosorbent assay strategy to measure neutralizing antibody response, 2 studies^30,31 used vesicular stomatitis virus glycoprotein pseudovirus expressed luciferase or Green fluorescent protein to detect neutralizing antibody responses. One study²⁶ used a high throughput alternative to the traditional neutralization assay that identified over 8 variants of receptor-binding domain antigens as supposed to the traditional 3 variants. When the percentage of patients demonstrating a neutralizing antibody response above study threshold after 3 doses was pooled, high heterogeneity (I² = 95.7%) was observed (Figure 4A). This heterogeneity was found to be because of one study,²⁶ which used an alternative to the traditional neutralization assay, and was removed as an outlier. With this study removed, the percentage of patients demonstrating a neutralizing antibody response above study threshold after 3 doses was 60.9% (95% CI, 53.2%-68.6%; I² = 0%; Figure 4B). In one RCT,⁷ neutralizing antibody positivity was observed in 60% of patients after 3 doses of mRNA-1273 versus 25% receiving placebo, respectively (risk ratio = 2.4; 95% CI, 1.5-4.0).

FIGURE 4.

Prevalence of neutralizing antibody response above study threshold after 3 doses of any mRNA SARS-CoV-2 vaccine in transplant recipients with outlier study (A, top panel) and without outlier study (B, middle panel) and prevalence of neutralizing antibody response above study threshold after 3 doses of any mRNA SARS-CoV-2 vaccine in transplant recipients according to SARS-CoV-2 variant and study (C, bottom panel). CI, confidence interval; SARS-CoV-2, severe acute respiratory syndrome coronavirus 2.

Two studies^26,27 quantified the percentage of transplant recipients mounting a neutralizing antibody response to SARS-CoV-2 VOC, such as the Alpha,^26,27 Beta,^26,27 Delta,^26,27 and Gamma²⁶ variants. Although the prevalence across studies varied widely (I² = 99.1%), the percentage of patients mounting a neutralizing antibody response above study threshold according to SARS-CoV-2 variant was similar within each study (Figure 4C). One RCT²⁷ found a smaller percentage of transplant recipients displayed a neutralizing antibody response above threshold after 2 doses of mRNA-1273 to all 3 variants versus wild-type virus (differences in proportions: Alpha, 19% [95% CI, 8%-31%]; Beta, 21% [95% CI, 10%-32%]; and Delta, 16% [95% CI, 4% to 28%]). After 3 doses of mRNA-1273 vaccine, the number of transplant recipients displaying a neutralizing antibody response above threshold was greater for each variant compared with placebo (differences in proportions: Alpha variant 45% [95% CI, 28%-61%], Beta variant 38% [95% CI, 21%-55%], and Delta variant 37% (95% CI, 20%-55%] via surrogate virus neutralization test and wild-type virus 35% [95% CI, 17%-51%], Alpha variant 36% [95% CI, 17%-51%], Beta variant 31% [95% CI, 15%-46%], and Delta variant 37% [95% CI, 19%-53%] via pseudovirus neutralization assay, with P < 0.001 for all comparisons).

Factors Associated With Humoral Response

Five studies investigated factors associated with humoral response.^{5,29,30,32,33} These studies could not be pooled as the multivariable logistic regression performed across these studies accounted and adjusted for different covariates. In 2 multivariable analyses,^29,30 mycophenolate use was associated with reduced odds of response across a cohort of kidney, liver, heart, and lung transplant recipients (OR = 0.28; 95% CI, 0.14-0.54; P < 0.001)²⁹ and in heart transplant recipients alone (OR = 0.1; 95% CI, 0.01-0.49; P = 0.01).³⁰ Additionally, in another multivariable analysis,²⁹ belatacept was associated with reduced odds of humoral response across a cohort of kidney, liver, heart, and lung transplant recipients (OR = 0.14; 95% CI, 0.04-0.46; P = 0.001), and, in one study⁵ of 62 belatacept-treated KTRs, only 4 patients (6.4%) developed anti–SARS-CoV-2 immunoglobulin G with low antibody titers (median 209, interquartile range 20–409 AU/mL).

Across multivariable analyses, lymphocyte count <1500 mm³ (OR = 0.31; 95% CI, 0.10-0.63; P = 0.004)³³ and lymphocyte count <1000 mm³ (OR = 0.20; 95% CI, 0.04-0.65; P = 0.02),²⁶ antiproliferatives (OR = 0.06; 95% CI, 0.01-0.41; P = 0.01),²⁶ triple immunosuppression (OR = 0.42; 95% CI, 0.21–0.86; P = 0.02),²⁹ and higher CRP (OR = 0.88; 95% CI, 0.79-0.96; P = 0.02)³⁰ were associated with nonhumoral response, whereas allograft function (OR = 1.03; 95% CI, 1.01-1.06; P = 0.02)³³ and higher eGFR (OR = 1.03; 95% CI, 1.01-1.06; P = 0.04)³⁰ were associated with an increased likelihood of achieving a positive antibody response in solid organ transplant recipients. Conflicting results were found with respect to age^29,30 and sex.^30,33 Finally, in a multivariable analysis of aHSCT recipients,³² only a B-cell count >0.25 g/L was associated with a humoral response (OR = 7.1; 95% CI, 1.5-34.1; P < 0.001).

Adverse Events

No serious (grade 3–4) adverse events or acute rejection episodes were reported across studies. Most common side effects included pain at injection site (62.4%), fatigue (18.5%), and headache (9.6%) with BNT162b2^26,30 and pain at injection site (76.7%), fatigue (43.3%), and myalgias (28.3%) with mRNA-1273 vaccine.⁷ One RCT⁷ found local and systematic events were more common after mRNA-1273 versus placebo.

Risk of Bias

All included noncomparative data, such as humoral response in a single group, were of good quality^5-7,25-33 (Table S4, SDC, http://links.lww.com/TP/C610), and all comparative data were of low risk of bias^7,27 (Table S5, SDC, http://links.lww.com/TP/C610). Funnel plot asymmetry was not detected (Figure S5, SDC, http://links.lww.com/TP/C610). When poor-quality studies³⁶ were included in the analysis, significant heterogeneity was detected (Figure S6, SDC, http://links.lww.com/TP/C610).^5-7,26-34

GRADE Assessments

Table S6 (SDC, http://links.lww.com/TP/C610) summarizes the certainty levels of all outcomes. Most outcomes had a low level (67%) or moderate level of certainty (17%). Most common reasons for decreasing certainty of the estimate included varying thresholds across studies and imprecision of the estimate.

DISCUSSION

Across solid organ and hematopoietic transplant recipients, this meta-analysis confirms 66.1% of transplant recipients overall and 45.9% of recipients without humoral response to 2 doses display a humoral response after 3 doses of mRNA SARS-CoV-2 vaccine, which did not significantly vary across transplant types. As humoral response has been shown to correlate with neutralizing antibody titers³⁸ and a relationship between neutralization level after SARS-CoV-2 vaccination and protection against COVID-19 has been demonstrated,⁹ this suggests humoral response to vaccination may be clinically relevant. Given approximately half of transplant recipients do not display humoral response after 2 doses,¹ a third dose of mRNA SARS-CoV-2 vaccine should be strongly considered. However, despite this, roughly a third of transplant recipients do not display humoral response after 3 doses, which suggests that this patient population remains at risk for infection.

Notably, humoral response after 3 doses was markedly attenuated with immunosuppression, as patients receiving triple immunosuppression, antiproliferatives, or belatacept had significantly reduced odds of developing a humoral response. Specifically, with belatacept use, only 6.4% of KTRs developed immunity. Thus, these patients seem to be at particular risk. Similarly, low lymphocyte counts were associated with markedly reduced response. Given the increased risk for progression to severe COVID-19 in high-risk patients who are vaccinated but are not expected to mount an adequate immune response, current guidelines recommend measures in addition to vaccination.³⁹ These measures include primary prevention, such as masking and social distancing, and secondary prevention, such as consideration of the administration of anti–SARS-CoV-2 monoclonal antibodies upon confirmation of SARS-CoV-2 infection.³⁹ Furthermore, in nonresponders and weak responders to 3 doses of vaccine, a fourth dose of SARS-CoV-2 vaccine may lead to an improved serological response vaccine.⁴⁰

Prevalence of humoral response was found to be lower in transplant recipients after 3 doses of mRNA-1273 versus BNT162b2. However, analysis of 1647 healthcare workers demonstrated higher antibody titers in participants vaccinated with 2 doses of mRNA-1273 compared with those vaccinated with BNT162b2.⁴¹ Given only 1 study was included in the mRNA-1273 group in our analysis and other analyses did not demonstrate a difference in efficacy between the 2 vaccines, it is likely that this result will be altered by future investigation.

Prevalence of humoral response did not significantly vary according to study thresholds. Given the wide variety of thresholds reported across studies, this result may change because more investigations are completed using similar thresholds to those already present in the literature. It is important to note that correlate of protection likely exists on a spectrum⁴² and that higher antibody levels are needed to achieve the same level of protection after vaccination compared with unvaccinated individuals with prior infection.⁴³ Thus, correlates of protection must be taken from vaccinated populations. Unfortunately, none of the studies in this review used thresholds that were based on clinical correlates of protection in vaccinated populations.⁴³ Individual studies⁷ used thresholds that conferred 50% protective neutralization in previous clinical studies of SARS-CoV-2–infected individuals. Other studies^6,32 used thresholds that correlated with in vitro neutralization of SARS-CoV-2. Therefore, the use of common, established thresholds that are proven to correlate with clinical protection in vaccinated individuals are required in future investigations of SARS-CoV-2 vaccine efficacy in transplant populations.

Prevalence of humoral response after 3 doses did not vary according to published correlates of protection for the wild type, alpha variant, and delta variant. This result suggests that mRNA SARS-CoV-2 vaccines provided transplant recipients protection from VOC, which is significant considering the higher risk of severe disease and hospitalization with certain variants, such as the Delta variant.⁴⁴ Future research is needed to establish correlates for protection for other VOC to guide vaccine administration and policy.⁴²

A third dose was associated with increased cellular and neutralizing antibody response. Specifically, the percentage of transplant recipients displaying a cellular response above study threshold after 3 doses was significantly higher compared with after 2 doses, with higher SARS-CoV-2–specific T-cell counts compared with placebo. Neutralizing antibody levels were also significantly increased after 3 doses of mRNA SARS-CoV-2 vaccine. As neutralization level is highly predictive of immune protection⁹ and T cells reduce viral loads and disease even when neutralizing antibody levels are low,¹⁰ these data suggest that a third dose may provide at least partial protection and reduce the likelihood of severe COVID-19 and infection in transplant recipients. Notably, the proportion of transplant recipients demonstrating a neutralizing antibody response above threshold to VOCs was higher after 3 doses compared with 2 doses, with a greater proportion of patients demonstrating a neutralizing antibody response to VOCs compared with placebo. Given neutralizing antibody positivity after 2 doses of mRNA-1273 vaccine was low for all 3 SARS-CoV-2 VOCs (Alpha, Beta, and Delta), these data suggest that a third dose may at least provide partial protection of transplant recipients to VOCs.

Across select cellular and neutralizing antibody outcomes, heterogeneity was observed. With respect to cellular response, this may be because of the use of different assays across studies. For example, the study³⁴ reporting the lowest prevalence measured a single cytokine whereas other studies^30,31 measured multiple cytokines in their cellular assays. In addition, heterogeneity was also observed with neutralizing antibody response to 3 doses. This heterogeneity was found to be because of one study,²⁶ which only allowed for 21 d between second and third dose administration versus >60 d for the other studies.^7,30 Given previous investigations have demonstrated that neutralizing antibody titers peak around 1 mo after infection^45,46 and decrease significantly thereafter,^45,47 high neutralizing antibody titers at the time of third dose administration (because of a short interval between second and third doses) may explain this study’s reduced neutralizing antibody titers after 3 doses. Additionally, this study²² used a commercially available high throughput alternative to the traditional neutralization assay that identified over 8 variants of receptor-binding domain antigens as supposed to the traditional 3 variants.²³ These differences could largely contribute to the heterogeneity in the results observed. Thus, standardization of cellular response assays, particularly with respect to measured cytokines, and neutralizing antibody assays could be considered. Future investigation of the cellular and neutralizing antibody response is required.

Interestingly, no consensus in relation to the tests used for cellular assays, neutralizing assays, or humoral assays was observed among the studies included in our analysis. Considering the assays differed according to measured cytokines (cellular assays) and had variable cutoffs (humoral and neutralizing assays), emphasis should be made toward standardizing the tools used to measure cellular and humoral responses. Alternatively, information should be provided on how the results from 1 assay correlate with other standard assays used. Such an effort would enable direct comparison and contrast of results across studies, which would minimize heterogeneity and allow for more robust recommendations.

A paucity of literature exists regarding humoral and cellular response in other populations after 3 doses of mRNA SARS-CoV-2 vaccine. Nonetheless, when compared with existing literature on other patient populations, transplant recipients seem to have a lower prevalence of humoral response. For example, in a cohort of patients on maintenance hemodialysis,⁴⁸ 92.4% of patients displayed humoral response after 3 doses compared with 66.1% of transplant recipients. Additionally, similar to transplant recipients with absent or minimal humoral response to 2 doses, small case series have demonstrated that a third dose of vaccine augments humoral response in patients with rheumatoid arthritis with absent or minimal humoral response to 2 doses.⁴⁹ Finally, meta-analyses^50,51 on the efficacy of 2 doses in patients with immune-mediated inflammatory diseases have demonstrated an association between attenuation of humoral response and anti-CD20 (rituximab) and anticytotoxic T lymphocyte–associated antigen therapies. This attenuation of humoral response has also been noted in transplant recipients receiving antimetabolites after 2 doses of vaccine.¹ As seen in our review, some immunosuppressive therapies were associated with reduced humoral response after 3 doses. Future investigation is required to determine interventions that can promote serological response in these populations, with some investigations already planned.⁵²

Limitations of this review include (1) the lack of literature reporting on cellular response and neutralizing antibodies and (2) the various assays and thresholds used across studies. Further studies are needed for standardization of SARS-CoV-2 serology and to correlate clinical protection with humoral and cellular response in this population.⁵³ Additionally, the inclusion of the poor-quality study resulted in significant heterogeneity. This study was judged to be of poor quality as, among other factors, this study did not report important data such as time since transplant and immunosuppressive regimens. As seen from our analyses, immunosuppressive regimens seem to greatly influence humoral response after 3 doses of vaccine. Thus, unreported factors, such as immunosuppressive regimen, may lead to the reduced humoral response reported in this poor-quality study. Finally, in this review, the duration of response and the effect of previous immunization, such as viral vector, were not able to be assessed.

In conclusion, a third dose SARS-CoV-2 mRNA vaccine should be strongly considered in transplant recipients. Future study and establishment of standardized assays and clinically relevant thresholds for humoral and cellular response are required. As this population seems to remain at risk for infection post 3 doses, investigations regarding the efficacy of future doses should be considered, and patients and healthcare providers should remain vigilant regarding exposure to infection.