Maintenance dialysis patients have higher coronavirus disease 2019 (COVID-19)-related mortality risk than the general population [1]. We and others have shown that patients have waning early antibody-mediated and blunted T cell-mediated immune responses to severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) vaccination [1, 2]. Optimizing the vaccination strategy in this population requires an understanding of the humoral and cellular immune response dynamics to SARS-CoV-2 vaccines, but immunogenicity data post-booster after primary COVID-19 vaccine cycle are scarce [3]. Here, we report follow-up data on the immune responses 6 months after primary COVID-19 vaccine cycle (T3) and 4 weeks post-booster (T4) following heterologous and homologous primary COVID-19 vaccine cycle SARS-CoV-2 vaccinations in adult patients receiving thrice weekly, in-center dialysis (hemodialysis and peritoneal dialysis) at the University Hospital Giessen and Marburg, Giessen, Germany [1].

We assessed anti-SARS-CoV-2 spike antibodies using a dot plot array (GenID, Strassberg, Germany) and chemiluminescent microparticle immunoassay (Anti-S AdviseDx anti-SARS-CoV-2 spike antibodies II, Abbott, Chicago, IL, USA), and T-cell responses by interferon (IFN)-γ and interleukin (IL)-2 peripheral blood leukocyte secretion upon SARS-CoV-2 glycoprotein stimulation (ELISpot assay, GenID; Supplementary Methods, Supplementary Table S1: study methods, statistical analysis, patients’ characteristics). The local human research ethics committee (AZ 126/21) approved this study and it complied with the Declaration of Helsinki tenets. All participants provided written informed consent before study enrollment.

Of the original cohort (n = 60), 47 patients (78.3%) were available for follow-up (T3: n = 42; T4: n = 46; five patients were transferred to other dialysis centers; six patients died from non-COVID-19-associated causes; two patients received boosters outside their dialysis center). Two patients had asymptomatic COVID-19 breakthrough infection despite complete primary COVID-19 vaccine cycle and therefore were only tested at T4 (Supplementary Table S2). The results of the timepoints T1–T2 around the primary COVID-19 vaccine cycle were recently published [1].

All patients received the mRNA-1273 mRNA-based vaccine booster (Moderna Biotech). Figure 1 depicts the humoral and cellular response dynamics 6 weeks (T2), and 6 months (T3) after primary COVID-19 vaccine cycle and 4 weeks (T4) after booster vaccination. The median anti-SARS-CoV-2 spike antibody levels (Abbott array) were significantly lower at T3 than T2 (501 [interquartile range, 134–1703] vs. 2240 [756–7687] arbitrary units [AU]/ml; P < 0.001), increasing markedly to 40,000 [6855–40,000] AU/ml post-booster (P < 0.001; Supplementary Tables S3, S4). No changes were observed for percent positivity status across T1–T4 (Fig. 1C).

Fig. 1
figure 1

Vaccine-induced anti-SARS-CoV-2 spike antibody detected using the Abbott array (A), GenID assay (B), and/or both (C), and SARS-CoV-2-specific T cell responses with secretion of IFN-γ (D), IL-2 (E), and/or both (F) at T2–T4. The figure depicts the cellular and humoral responses at 6 weeks (T2), 6 months (T3) after basic vaccination, and at 4 weeks (T4) post-booster. The humoral response level (as determined by the Abbott array and GenID assay) was lower at T3 compared to T2 (P < 0.001) but increased post-booster (P < 0.001). There was no reduction in the IFN-γ response between T2 to T3 (P = 0.552) while the SARS-CoV-2-specific IL-2 response was reduced between both timepoints (P = 0.023). No increase in cellular response (IL-2 or IFN-γ) was observed post-booster (p = NS). A logarithmic scale was used on the y-axis in panel A, D, and E. Due to the log scale, anti-SARS-CoV-2 spike antibody (Abbott array), IFN-γ, and IL-2 levels of zero are not displayed. The dashed horizontal lines indicate the cut-off for positivity (reactive; i.e., IgG > 50 AU/ml [Abbott array] and > 16% [GenID assay], IFN-γ and IL-2: SI ≥ 7); the area between the horizontal lines indicates the borderline zone used in each GenID assay. Bold values denote statistical significance at the P < 0.05 level. AU arbitrary unit, IFN-γ interferon-γ, IgG immunoglobulin G, IL-2 interleukin-2, SARS-CoV-2 severe acute respiratory syndrome coronavirus type 2, NS not significant, T2 timepoint 2, T3 timepoint 3, T4 timepoint 4

Full size image

The median IL-2 stimulation index levels were lower at T3 than T2 (P = 0.023) but not the IFN-γ stimulation index levels (P = 0.552) between both timepoints (Fig. 1D–E, Supplementary Table S3). Notably, IFN-γ stimulation index levels were higher at T4 than T2. No changes were observed when comparing the percent reactive pattern of the IFN-γ and/or IL-2 ELISpot assays across T1–T4, but the results were flawed due to the high number of invalid samples (Fig. 1F).

The GenID assay demonstrated that patients with IFN-γ-producing T cells had higher anti-SARS-CoV-2 spike antibody levels at T3 (P = 0.028, n = 30) but not the Abbott array (P = 0.08; n = 28). At T4, there was no significant difference for either assay (Abbott array: P = 0.51, n = 17; GenID assay: P = 0.442, n = 17). IL-2 could not be analyzed due to the low numbers on the reactive side at T3 (n = 1) and T4 (n = 3).

Patients with COVID-19 history had sustained higher anti-SARS-CoV-2 spike antibody levels (Abbott array) compared to infection-naïve patients at T2 (n = 5 vs. 53, respectively, total number = 58) (P < 0.001) and T3 (n = 5 vs. 35, respectively, total number = 40) (P = 0.002; Supplementary Table S5), although the booster conferred median IgG levels reaching the upper detection limit of 40,000 AU/ml in both groups at T4 (n = 6 vs. 36, respectively, total number = 42). Patients with COVID-19 history also had higher SARS-CoV-2-specific IFN-γ levels at T2 (P < 0.001), but not IL-2 (P = 0.07). No differences were seen in the IFN-γ SI levels at T3 (P = 0.252) and T4 (P = 0.299) between both groups (Supplementary Table S6). Given the high number of invalid samples of patients with COVID-19 history, the T3 and T4 IL-2 immune responses could not be analyzed.

Our results indicate a robust humoral immune response 6 months following primary COVID-19 vaccine cycle (> 90%), which is consistent with previous reports involving hemodialysis patients and healthy controls [3, 4]. However, while primary COVID-19 vaccine cycle resulted in markedly high anti-SARS-CoV-2 spike antibody levels (levels were highest in patients with previous COVID-19), the humoral response waned significantly within 6 months. IgG seropositivity, defined by commercially available tests, may overestimate the effectiveness of vaccine-induced humoral immunity, as the cutoff value that correlates with protection against SARS-CoV-2 infection is unknown. In contrast, we observed a sustained weak cellular immune response post-booster, although IFN-γ stimulation index levels increased significantly. Therefore, in line with previous works [4], antibody presence may not automatically correlate with functional cellular immunity, which is likely an important component in long-term protection against SARS-CoV-2. We and others have previously shown that cytokine induction during primary infection is associated with preferential induction of T cells producing IL-2, whereas reactivations are associated with T cells producing IFN [5]. This may also be applicable to booster vaccinations, as shown in the present study. Overall, our data indicate progressive waning of humoral immunity and a sustained weak cellular immune response within 6 months; the booster vaccination is able to substantially increase humoral immunity again; the emergence of SARS-CoV-2 variants with high potential for immune evasion may necessitate a further booster dose 4–6 months after the previous booster vaccination in dialysis patients.