Immune Responses after a Third Dose of mRNA Vaccine Differ in Virus-Naive versus SARS-CoV-2– Recovered Dialysis Patients

Abstract

Visual Abstract

Background and objectives 

After two doses of mRNA vaccine against severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), patients on dialysis show a defective humoral response, but a third dose could increase anti–SARS-CoV-2 spike IgG titers. Responses could be different in virus-naive and SARS-CoV-2–recovered patients on dialysis. However, characterization of memory B cell response after three doses is lacking.

Design, setting, participants, & measurements 

We evaluated the dynamics of antireceptor binding domain IgG titers and antireceptor binding domain memory B cells until 6 months after two and three doses (administered within 6 months after the second dose) of mRNA vaccine in SARS-CoV-2–recovered and virus-naive dialysis populations. Results were analyzed by ordinary one-way ANOVA, the Kruskal–Wallis test, or the Wilcoxon matched-pairs test as appropriate.

Results 

In total, 108 individuals (59 patients on dialysis and 49 controls) were included. In virus-naive patients on dialysis, antireceptor binding domain IgG response was quantitatively lower after two doses compared with healthy controls, but IgG titers increased by three-fold after three doses (P=0.008). In SARS-CoV-2–recovered patients on dialysis, antireceptor binding domain IgG titers after two doses were significantly higher compared with virus-naive patients on dialysis but did not significantly increase after a third dose. Regarding memory B cell response, we detected receptor binding domain–specific memory B cells at similar proportions in virus-naive patients on dialysis and vaccinated controls after two doses. Moreover, a strong receptor binding domain–specific memory B cell expansion was observed after the third dose in virus-naive patients on dialysis (5.5-fold; P<0.001). However, in SARS-CoV-2–recovered patients on dialysis, antireceptor binding domain memory B cells remained unchanged after the third dose.

Conclusions 

The third dose of mRNA vaccine given within 6 months after the second dose boosts serologic and memory response in virus-naive patients but not in SARS-CoV-2–recovered patients on dialysis.

Clinical Trial registry name and registration number: 

COVID-19: SARS-CoV-2 Specific Memory B and T-CD4+ Cells (MEMO-COV2), NCT04402892

Introduction

Patients with kidney failure requiring hemodialysis have a high mortality rate after infection with severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) (^1–3). Additionally, in-center hemodialysis is associated with a higher risk of contact with health professionals or other high-risk patients, which may explain the elevated incidence of SARS-CoV-2 infection observed in patients on dialysis (⁴,⁵). Consequently, the dialysis population appears as a paradigm of a high-risk population in the SARS-CoV-2 pandemic. The rapid vaccine rollout had a major role in controlling infection spread (⁶), but patients on dialysis have been largely excluded from vaccine trials for safety reasons. To date, characterization of SARS-CoV-2 vaccine response in patients on dialysis is mainly on the basis of serologic response, showing an 80%–95% rate of positive antispike IgG serology after mRNA vaccination (⁷), although antibody levels appeared reduced when compared with the nondialysis population (^7–10). The analyses of vaccine efficacy in patients on dialysis suggest a clinical efficacy against coronavirus 2019 (COVID-19) (¹⁰,¹¹), but this efficacy could be lower than in the general population because severe forms are not rare even after the complete vaccine scheme (two doses) (¹²). In parallel to the rapid serum IgG response, vaccination induces the progressive generation of memory B cells, which continuously acquire somatic mutations in their variable region genes to improve affinity and constitute another layer of immune protection. After antigenic re-exposure, memory B cells subsequently drive the recall response by differentiating into new antibody-secreting cells, providing high-affinity protective antibodies. In healthy individuals, although the antibody level wanes over time, memory B cells continued to increase in frequency and to mature between 3 and 6 months postvaccination (¹³,¹⁴) (R.R. Goel et al., unpublished data), and thus, they constitute a key secondary layer of protection. Conversely, defective early humoral responses against SARS-CoV-2 have been reported in patients on dialysis and kidney transplant recipients (¹⁵).

The frailty of the dialysis population and concerns regarding long-term vaccine response (¹⁶) and clinical efficacy (¹²) have led to the recommendation of a third dose of vaccine since April 2021 for patients on dialysis in France. Several studies have shown an increase of antispike IgG response in virus-naive patients on dialysis after a third dose of mRNA vaccine (¹²,^17–19), but the dynamics and persistence of the SARS-CoV-2–specific memory B cell pool after vaccination have not yet been characterized in the dialysis population. Similarly, the effect of a third vaccine dose in SARS-CoV-2–recovered individuals remains unclear, although patients who recovered from COVID-19 may now represent up to 20%–30% of the dialysis population in some areas (²⁰). Detailed analyses of the vaccine response in this subgroup could have major importance in understanding the immunologic response after repeated antigen exposure.

Here, we characterized serologic and memory B cell response against the receptor-binding domain (RBD) region of the SARS-CoV-2 spike protein after mRNA vaccination in previously SARS-CoV-2–infected (SARS-CoV-2–recovered) and virus-naive patients on hemodialysis.

Materials and Methods

Patients on Dialysis

Fifty-nine patients on dialysis who received the BNT162b2 mRNA vaccine from three dialysis centers in the Ile-de-France region (Centre Hospitalier Universitaire Henri Mondor, Assistance Publique-Hôpitaux de Paris, Creteil; Hôpital Privé Nord Parisien, Sarcelles; and Clinique de Villeneuve Saint Georges–Ramsay Sante, Villeneuve Saint Georges) were included in this study. Among them, 33 were naive to SARS-CoV-2 (virus naive), and 26 recovered from SARS-CoV-2 (SARS-CoV-2 recovered). Six SARS-CoV-2–recovered patients on dialysis had severe COVID-19 (hospitalized COVID-19 requiring oxygen), and 22 had mild COVID-19 (ambulatory disease). Median (minimum to maximum) time from COVID-19 diagnosis to first vaccine injection was 8 (2–11) months.

All patients on dialysis received at least two doses of the BNT162b2 mRNA vaccine at a mean time of 27.6 (±1.7) days after the first dose. Anti-RBD IgG titers and RBD-specific memory B cells were assessed 1–2 months after the second dose (anti-RBD IgG titers: mean 36 [SD±22] days after the second dose and RBD-specific memory B cells: mean 66 [SD±8] days after the second dose).

Among virus-naive patients on dialysis, 26 patients received a third dose at a median 145 (interquartile range [IQR], 70–153) days after the second dose and were sampled at a median of 50 (IQR, 45.5–50) days after the third dose.

Among SARS-CoV-2–recovered patients on dialysis, 13 received a third dose at a median of 145 (IQR, 70–154) days after the second dose and were sampled at a median of 49 (IQR, 42–50) days after the third injection. Five patients with only two vaccine injections were also sampled 6 months after the second dose (mean time 200 [SD±4] days from the second dose). Only one patient developed SARS-CoV-2 infection during follow-up. Vaccine policies were left to the discretion of the caring physician of the different centers, but no difference was observed in patients’ characteristics according to the number of vaccine doses.

The COVAC-HD study was approved by the ethical committee ELSAN institutional review board (2021-04-ATTIAS-01) and was performed in accordance with French law. Written informed consent was obtained from all participants.

Vaccinated Controls

Forty-nine individuals with no kidney disease (vaccinated controls) were included from the original MEMOCOV2 cohort (institutional review board 2018-A01610-55), including 23 healthy health care workers naive to SARS-CoV-2 (no history of COVID-19 and negative IgG antinucleocapsid and/or spike) and 26 patients recovered from mild or severe COVID-19 (SARS-CoV-2 recovered).

All vaccinated controls received the BNT162b2 mRNA vaccine. Virus-naive vaccinated controls received two doses at a mean of 27.7 days (SD±1.8 days) apart as part of the French vaccination campaign. SARS-CoV-2–recovered vaccinated controls received only one dose in line with French guidelines, except for three patients who received two doses. First injection was performed a mean of 308 days (SD±45 days) after the infection. Both cohorts were sampled 1–2 months after the first dose (SARS-CoV-2 recovered) or the second dose (SARS-CoV-2 recovered; mean ± SD: 64.05±13.3 and 63.3±9.0 days for naive, respectively).

Vaccinated control individuals were recruited at Henri Mondor University Hospital (Assistance Publique-Hôpitaux de Paris) between March and April 2021. The COVID-19: SARS-CoV-2 Specific Memory B and T-CD4+ Cells study (NCT04402892) was approved by the ethical committee of Ile-de-France VI (no. 40–20 HPS) and was performed in accordance with French law. Written informed consent was obtained from all participants.

Patients’ Characteristics

SARS-CoV-2 infection was defined as confirmed RT-PCR on nasal swab or clinical presentation associated with typical aspect on computed tomography scan and/or serologic evidence. SARS-CoV-2–naive individuals were defined as no history of COVID-19 and negative IgG antinucleocapsid (and/or spike). Immunocompromised patients were defined as patients with former kidney transplantation, current immunosuppressive treatment/chemotherapy or immunosuppressive treatment/chemotherapy discontinued for <6 months, or HIV infection (with CD4⁺ T cells below 500/mm³).

Statistical Analyses

Ordinary one-way ANOVA and Kruskal–Wallis tests were used to compare continuous variables as appropriate (indicated in Figures 1 and 2). Wilcoxon matched-pairs signed rank tests for longitudinal analysis were used. A P value =0.05 was considered statistically significant. Statistical analyses involved use of GraphPad Prism 9.0 (La Jolla, CA).

Figure 1.: Humoral and memory B cell response to severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) receptor-binding domain (RBD) in virus-naive and SARS-CoV-2–recovered patients on dialysis after mRNA vaccination. (A) Antibody response. Median ± interquartile range (IQR) anti−SARS-CoV-2 RBD serum IgG titers measured by ELISA in virus-naive (left panel; white dots) and SARS-CoV-2–recovered individuals (right panel; mild coronavirus disease 2019 [COVID-19]: orange dots; severe COVID-19: red dots). White bars represent virus-naive (n=23) or SARS-CoV-2–recovered (n=26) healthy controls 1–2 months after two doses or one dose of mRNA vaccine, respectively (#). Light blue bars represent virus-naive (n=20) and SARS-CoV-2–recovered (n=20) patients on dialysis 1–2 months after a second dose of mRNA vaccine. Dark blue bars represent virus-naive (n=25) and SARS-CoV-2–recovered (n=13) patients on dialysis 1–2 months (median of 50 days) after a third dose of mRNA vaccine. (B) Antibody response. Longitudinal evolution of anti–SARS-CoV-2 RBD IgG titers after a third dose of mRNA vaccine in virus-naive (n=13) and SARS-CoV-2–recovered (n=7) patients on dialysis (median time from second dose to third dose: 153.5 [IQR, 145–155] days) and 6 months after a second dose in SARS-CoV-2–recovered (n=6) patients on dialysis. (C) Memory B cell response. Median ± IQR frequencies of SARS-CoV-2 RBD-specific cells in live CD19⁺IgD⁻CD27⁺CD38^int/− memory B cells by flow cytometry in virus-naive and SARS-CoV-2–recovered individuals. (D) Memory B cell response. Frequencies of SARS-CoV-2 RBD-specific memory B cells after a third dose of mRNA vaccine in virus-naive (n=14) and SARS-CoV-2–recovered (n=7) patients on dialysis (median time from second dose to third dose: 153.5 [IQR, 145–155] days) and 6 months after a second dose in SARS-CoV-2–recovered (n=5) patients on dialysis. Kruskal–Wallis test with multiple comparisons between groups for (A) and (C) and Wilcoxon matched-pairs signed rank test for (B) and (D). Bars indicate median ± IQR. Values equal to zero were converted to 0.01 for logarithmic scale. **P<0.01; ***P<0.001; ****P<0.0001. 6M, 6 months; ns, nonsignificant; S, spike; A.U, arbitrary unit.

Figure 2.: Characterization of the anti-RBD B cell response. (A) Median ± IQR of frequencies of SARS-CoV-2 RBD-specific cells in live CD19⁺IgD⁻CD27⁻CD38^int/− DN B cells by flow cytometry in virus-naive (left panel; white dots) and SARS-CoV-2–recovered individuals (right panel; mild COVID-19: orange dots; severe COVID-19: red dots). White bars represent virus-naive (n=23) or SARS-CoV-2–recovered (n=23) healthy controls 1–2 months after two doses or one dose of mRNA vaccine, respectively (#). Light blue bars represent virus-naive (n=19) and SARS-CoV-2–recovered (n=17) patients on dialysis 1–2 months after a second dose of mRNA vaccine. Dark blue bars represent virus-naive (n=23) and SARS-CoV-2–recovered (n=13) patients on dialysis 1–2 months (median 50 days) after a third dose of mRNA vaccine. (B) Median ± IQR of frequencies of SARS-CoV-2 RBD-specific cells displaying a DN2 (CD19⁺CD27⁻IgD⁻CD11c⁺CD21⁻) or a DN1 (CD19⁺CD27⁻IgD⁻CD11c⁻CD21⁺) phenotype in virus-naive (n=9) and SARS-CoV-2–recovered (=18) healthy controls 1–2 months after two doses or one dose of mRNA vaccine, respectively; in virus-naive (n=9) and SARS-CoV-2–recovered (n=15) patients on dialysis 1–2 months after a second dose of mRNA vaccine; and in virus-naive (n=17) and SARS-CoV-2–recovered (n=9) patients on dialysis 1–2 months (median of 50 days) after a third dose of mRNA vaccine. We performed repeated measures mixed effects model analysis with multiple comparisons between time points for dialysis groups and Kruskal–Wallis with multiple comparisons between donor groups inside each time point and between vaccinated controls groups. Bars indicate median ± IQR. Values equal to zero were converted to 0.01 for logarithmic scale. *P<0.05; **P<0.01; ***P<0.001; ****P<0.0001. DN, double negative; S, spike; AU, arbitrary unit.

Detailed methods are provided in Supplemental Material.

Results

In total, 108 patients were followed in this study, including 59 patients on dialysis from three centers of the Paris area (Table 1). All patients received two doses (1 month apart) of the BNT162b2 vaccine (Pfizer/BioNTech) and then a third injection between 2 and 6 months after the first dose. We previously characterized the longitudinal evolution and maturation of SARS-CoV-2–responding B cells after mRNA vaccine in healthy individuals (²¹). Forty-nine individuals from this original cohort (virus naive or SARS-CoV-2 recovered) were included as controls and had all been sampled 1–2 months after the boost injection (second dose in virus-naive individuals and first dose in SARS-CoV-2–recovered individuals). Details of the study design and methodology are provided in Supplemental Material (Supplemental Figure 1).

Table 1. - Clinical characteristics of vaccinated controls and patients on dialysis according to their severe acute respiratory syndrome coronavirus 2 infection status

Variable	Vaccinated Controls		Patients on Dialysis
	Naive, n=23	Recovered, n=26	Naive, n=33	Recovered, n=26
Men, n (%)	11 (48)	9 (35)	23 (70)	17 (65)
Age, yr, median [IQR]	45 [33–54]	45 [36–59]	71 [67–77]	70 [65–73]
Comorbidities
Diabetes mellitus, n (%)	—	2 (8)	16 (48.5)	11 (42)
Hypertension, n (%)	—	3 (12)	32 (97)	25 (96)
Immunodepression, n (%)	—	0 (0)	5 (15)	2 (8)
Former kidney transplantation			1 (3)	1 (4)
Immunosuppressive therapy			2 (6)	1 (4)
Cancer with chemotherapy			1 (3)
HIV			1 (3)
Dialysis vintage, yr, median [IQR]			5.2 [1.9–6.4]	3.95 [2.7–5]
COVID-19 characteristic, n (%)
Mild presentation	—	20 (77)	—	19 (73)
Severe presentation	—	6 (23)	—	7 (27)
mRNA vaccine details
Sample available after 2 doses			21 (64)	20 (77)
Sample available after 3 doses			26 (79)	13 (50)
Sample available after 2 doses + 6 mo postboost				6 (23)

IQR, interquartile range; COVID-19, coronavirus disease 2019; —, no patients (0) had diabetes, hypertension, or immunodepression in naive vaccinated controls.

We first measured the pre- and postboost evolution of IgG serum titers against the RBD in virus-naive and SARS-CoV-2–recovered patients on dialysis (Supplemental Tables 2–5). We focused on the RBD region of the spike because anti-RBD antibodies contribute to the majority of neutralizing antibodies against SARS-CoV-2. After the second dose of mRNA vaccine, anti-RBD IgG response was quantitatively lower in virus-naive patients on dialysis compared with virus-naive controls (Figure 1A) as previously reported (although the difference did not reach statistical significance in this cohort) (⁸,⁹). Strikingly, a third dose elicited a strong recall response, with anti-RBD IgG titers increasing, on average, by three-fold compared with titers postsecond dose (P=0.008) (Figure 1, A and B). These levels were even higher than those of virus-naive nondialysis controls after two doses (P<0.001). In contrast, SARS-CoV-2–recovered patients on dialysis had reached elevated anti-RBD IgG serum titers after their second vaccine dose. These titers were higher than those of SARS-CoV-2–recovered vaccinated controls (who received only one dose according to French vaccine policy; P=0.005) (Figure 1A). Moreover, in SARS-CoV-2–recovered patients on dialysis, the third dose did not lead to any significant increase in serum anti-RBD IgG titers (Figure 1, A and B). We conclude that the third vaccine dose, when administered within 6 months, induced a strong recall response in virus-naive patients on dialysis but had no significant effect on anti-RBD IgG titers in SARS-CoV-2–recovered patients on dialysis.

To further explore these two distinct patterns of response, we next studied RBD-specific CD27⁺IgD⁻ memory B cells by flow cytometry as previously described (Supplemental Figure 2A, Supplemental Tables 1–5) (²¹).

We first analyzed the RBD-specific memory B cell frequencies after the second dose of mRNA vaccine in virus-naive individuals. Interestingly, we detected RBD-specific memory B cells at similar proportions in patients on dialysis and vaccinated controls, suggesting that the settlement of the memory B cell pool was not impaired in the virus-naive dialysis population. After a third dose of mRNA vaccine, we further observed a strong RBD-specific memory B cell expansion in virus-naive patients on dialysis with a 5.5-fold increase as compared with their frequencies after the second injection (P<0.001) (Figure 1, C and D). We next studied SARS-CoV-2–recovered patients on dialysis. These patients already displayed robustly expanded RBD-specific memory B cells after two doses of mRNA vaccine, reaching frequencies similar to SARS-CoV-2–recovered nondialysis controls and significantly higher than virus-naive patients on dialysis and nondialysis controls (Figure 1C). These levels after two doses were statistically higher in patients with prolonged delay between infection and first vaccination (>4 months) (Supplemental Figure 3). Age was associated with a reduced memory B cell expansion in naive patients on dialysis but not in SARS-CoV-2–recovered patients (Supplemental Figures 3 and 4). Moreover, after the third dose, no further increase of RBD-specific memory B cells was observed in SARS-CoV-2–recovered patients on dialysis. Furthermore, RBD-specific memory B cell frequencies were similar in SARS-CoV-2–recovered and virus-naive controls (Figure 1, C and D). Lastly, we took advantage of serial samples in selected patients to show that frequencies of RBD-specific memory B cells remain stable in SARS-CoV-2–recovered patients on dialysis up to 6 months after the second dose. We conclude that the third vaccine dose (administered within 6 months after the second dose) induced an expansion of RBD-specific memory B cells in virus-naive patients on dialysis but had no significant effect on RBD-specific memory B cells in SARS-CoV-2–recovered patients on dialysis.

Finally, to better characterize the generated anti-RBD–specific B cell subsets in patients on dialysis, we performed multiparametric FACS analysis using a flow panel that included seven surface markers (CD19, CD21, CD27, CD11c⁺, CD38, CD71, IgD) (Supplemental Figure 2, B and C). The RBD-specific double-negative IgD⁻CD27⁻ population exhibits an expansion in virus-naive patients on dialysis after the third dose (Figure 2A), albeit with no difference in the repartition of atypical memory B cells (DN2; IgD⁻CD27^+/−CD11c⁺) and antibody-secreting cells precursors (DN1; IgD⁻CD27⁻) (Figure 2B). No further increase of RBD-specific double-negative IgD⁻CD27 was observed in SARS-CoV-2–recovered patients on dialysis between the second and third doses (Figure 2B).

Discussion

In this study, we describe antibody and memory B cell response after a third dose of mRNA vaccine in infection-naive and recovered patients on dialysis. Previous reports have suggested that the humoral response was impaired in patients on dialysis (⁹,¹⁵,²²). We provide here a more complex view showing that virus-naive and SARS-CoV-2–recovered patients on dialysis display different profiles of immune response. Notably, a third dose of mRNA vaccine administered within 6 months after the second dose strongly boosted both serologic and cellular memory responses in virus-naive patients on dialysis. However, this boost was not observed in SARS-CoV-2–recovered patients previously vaccinated with two doses. As previously reported in larger studies (⁸,⁹), we show that the anti-RBD IgG response remained quantitatively lower (but nonsignificantly in our cohort) in virus-naive patients on dialysis compared with virus-naive healthy controls after two doses of mRNA vaccine. We demonstrate for the first time that RBD-specific memory B cells appeared settled at similar levels after two doses of mRNA vaccine in virus-naive patients on dialysis and nondialysis controls. Furthermore, these memory B cells were efficiently recalled by a third dose of mRNA vaccine as illustrated by the strong anti-RBD antibody response, exceeding the postsecond dose response in controls, and the high memory B cell frequencies, similar to those of SARS-CoV-2–recovered patients on dialysis. Of note, regarding the current third dose policy proposed to the general population, our data suggest a high benefit of this strategy in virus-naive individuals because this recall could be even more pronounced in the nondialysis population (¹⁵). Conversely, in SARS-CoV-2–recovered patients on dialysis, the third dose administered within 6 months did not further increase the anti-RBD IgG titers or the number of RBD-specific memory B cells, suggesting a limited benefit of such a rapid third dose in previously infected patients. Consequently, our results suggest that virus-naive patients on dialysis do not have a maximal immune response (antibodies as well as memory B cells) after two doses, and therefore, a third dose can have a booster effect. Conversely, in SARS-CoV-2–recovered patients, the immune response is already maximally activated after infection and two doses, and therefore, a third dose given within 6 months may have no additional booster effect. However, a potential benefit of a delayed third dose in SARS-CoV-2–recovered patients and whether the same results would be found if the third dose was given later remain to be studied. In line with this hypothesis, a better response after two doses was observed in patients with a prolonged interval between infection and vaccination (Supplemental Figure 3). Moreover, even in the absence of an anti-S increase, we cannot rule out a benefit of the third dose in terms of antibody affinity and neutralizing potential. Lastly, how the memory B cell pool generated after vaccination will evolve remains to be studied.

In addition to serum IgG, memory B cells constitute another layer of immune protection. We and others have shown that mRNA vaccination generated a broad and diverse repertoire of RBD-specific memory B cells that can recognize SARS-CoV-2 variants of concern (¹³). In this study, we were able to characterize this response in patients on dialysis up to 6 months after vaccination. We found that patients on dialysis (either SARS-CoV-2 naive or recovered) achieve a quantitatively normal RBD memory B cell response after two or three doses of mRNA vaccine in SARS-CoV-2–recovered and virus-naive patients, respectively. Moreover, we took advantage of a prolonged longitudinal follow-up to show that RBD memory B cell response remains stable up to 6 months after the second dose in SARS-CoV-2–recovered patients on dialysis, as recently reported in SARS-CoV-2–recovered patients not on dialysis (²³). Consequently, our data suggest that despite their immunocompromised status, patients on dialysis could settle a memory B cell response close to that of the general population. Whether the underlying repertoire of these RBD-specific memory B cells of patients on dialysis is as diverse and broadly neutralizing as in vaccinated healthy individuals remains an open question.

The current Omicron variant of concern is associated with very high transmissibility. From an epidemiologic point of view, our data could have major consequences in implementing future vaccination campaigns because the frequency of previous SARS-CoV-2 infection is obviously bound to increase in the dialysis population. By including the delay with (1) previous SARS-CoV-2 infection and (2) second dose in prioritization decisions, our results will lead to a better stratification of future vaccine policies. Notably, even though patients on dialysis seem to establish a memory B cell response similar to that of the general population, the characterization of the postvaccine B cell response in the other immunocompromised populations still needs to be studied.

The Omicron variant has become dominant in most countries. Omicron RBD harbors 15 mutations compared with the ancestral Hu-1 RBD analyzed in this study, and studies have shown decreased sera neutralization against this variant of concern. Although this limits the interpretation of our work, we recently showed that despite the strong immune escape of Omicron, the memory B cell pool generated after two doses of mRNA vaccines contains clones that maintained their neutralization potency against Omicron (²⁴). This is in line with numerous reports showing increased neutralizing antibodies titers against Omicron after a third dose of mRNA vaccine, reflecting the recruitment of these memory B cells (²⁵,²⁶).

Our study has several limitations. Obviously, controls and patients on dialysis have different characteristics, which limit their comparison and suggest that dialysis status may not solely explain differences in immune responses. The relatively small size and the limited follow-up of the cohorts do not allow us to evaluate the role of comorbidities (such as immunosuppression) in patients on dialysis or predict long-term response to mRNA vaccine and its persistence, although our results precisely characterize memory B cell response up to 6 months after vaccination. Also, longitudinal follow-up was not available for all patients. Given the retrospective design of the study, different timings between the second and third doses and SARS-CoV-2 infection and vaccination were possible. However, similarly to our whole cohort, we were able to show in selected patients with serial samples the increase of memory B cell response in virus-naive patients and the stability of this response in SARS-CoV-2–recovered patients (even without a third dose) (Figure 1, C and D). Moreover, our analysis in infection-experienced patients on dialysis may be altered by survivor bias, limiting the generalization of our observations to other at-risk populations. We cannot rule out that there may have been misclassification of previously infected status, at least of asymptomatic cases. Lastly, we evaluated responses to only BNT162B2 and not to other vaccines (including other mRNA vaccines or other vaccine platforms).

In conclusion, our study shows that previous SARS-CoV-2 infection strongly affects vaccine responses in patients on dialysis. The third dose of mRNA vaccine administered within 6 months after the second dose boosts the serologic and memory response in virus-naive patients on dialysis, but this boost was not observed in SARS-CoV-2–recovered patients on dialysis. These results suggest that the timing of the third dose and a previous SARS-CoV-2 infection have a critical effect in vaccine response of patients on dialysis and should be considered in future vaccine policies of this population.

Disclosures

K. El Karoui reports research funding from Amgen, Otsuka, and Sanofi and honoraria from Alexion, AstraZeneca, and Otsuka. S. Fourati reports consultancy agreements with Abbott Diagnostics, GeneXpert, GlaxoSmithKline (GSK), and MSD; honoraria from Abbott Diagnostics, GeneXpert, GSK, and MSD; serving in an advisory or leadership role for Abbott Diagnostics, GeneXpert, GSK, and MSD; and speakers bureau for Abbott Diagnostics, GeneXpert, GSK, and MSD. M. Mahévas reports research funding from GSK and Takeda and honoraria from Amgen and Novartis. All remaining authors have nothing to disclose.

Funding

None.