Real-world assessment of immunogenicity in immunocompromised individuals following SARS-CoV-2 mRNA vaccination: a two-year follow-up of the prospective clinical trial COVAXID

Abstract

Background: Immunocompromised patients with primary and secondary immunodeficiencies have shown impaired responses to SARS-CoV-2 mRNA vaccines, necessitating recommendations for additional booster doses. However, longitudinal data reflecting the real-world impact of such recommendations remains limited.

Methods: This study represents a two-year follow-up of the COVAXID clinical trial, where 364 of the original 539 subjects consented to participate. 355 individuals provided blood samples for evaluation of binding antibody (Ab) titers and pseudo-neutralisation capacity against both the ancestral SARS-CoV-2 strain and prevalent Omicron variants. T cell responses were assessed in a subset of these individuals. A multivariate analysis determined the correlation between Ab responses and the number of vaccine doses received, documented infection events, immunoglobulin replacement therapy (IGRT), and specific immunosuppressive drugs. The original COVAXID clinical trial was registered in EudraCT (2021-000175-37) and clinicaltrials.gov (NCT04780659).

Findings: Several of the patient groups that responded poorly to the initial primary vaccine schedule and early booster doses presented with stronger immunogenicity-related responses including binding Ab titres and pseudo-neutralisation at the 18- and 24-month sampling time point. Responses correlated positively with the number of vaccine doses and infection. The vaccine response was blunted by an immunosuppressive state due to the underlying specific disease and/or to specific immunosuppressive treatment.

Interpretation: The study results highlight the importance of continuous SARS-CoV-2 vaccine booster doses in building up and sustaining Ab responses in specific immunocompromised patient populations.

Funding: The present studies were supported by the European Research Council, Karolinska Institutet, Knut and Alice Wallenberg Foundation, Nordstjernan AB, Region Stockholm, and the Swedish Research Council.

Keywords: COVID-19; Chronic lymphocytic leukemia; Clinical study; HIV; Hematopoietic stem cell transplantation; Primary immunodeficiency disease; SARS-CoV-2; Solid organ transplantation; mRNA vaccine.

Fig. 1

Dynamics of antibody titres in the COVAXID cohort. (A) Prevailing major SARS-CoV-2 subvariants in Sweden during the study period. (B) Dynamics of Spike Wu-Hu.1 Ab titres (geometric mean with 95% CI, shaded range) for each subgroup. The vertical dotted line represents the 1-year follow-up sample timepoint. (C) Bar plots showing Spike Wu-Hu.1 Ab titres at 3, 6, 9, 12, 18 and 24 months at a study group level. Statistical tests were performed using Mann–Whitney, and Bonferroni correction for multiple comparisons, using the 24-month timepoint as reference. (D) Seroconversion rates over time in each subgroup as defined by Spike receptor-binding domain (RBD) titres ≥0.8 AU/ml. The star annotation (∗) indicates statistical significance at a p-value threshold of 0.05 (or ∗∗ for p < 0.01, ∗∗∗ for p < 0.001, ∗∗∗∗ for p < 0.0001).

Fig. 2

Antibody titres and pseudo-neutralising capacity of SARS-CoV-2 subvariants. Correlation matrices assessing (A) Ab titres and (B) pseudo-neutralising capacity of SARS-CoV-2 subvariants at the 24-month sample timepoint (Pearson correlation). (C) Bar plots showing neutralising capacity of four SARS-CoV-2 subvariants at the 24 months sampling point based on virus subvariant (upper panel) and study group (lower panel). Statistical tests were performed using Mann–Whitney, and Bonferroni correction for multiple comparisons, using healthy controls (HC) and Wu-Hu.1 (WT virus), respectively, as reference. (D) Line plot showing neutralising capacity dynamics of Spike Wu-Hu.1 and Spike BA.1 over time (geometric mean with 95% CI, shaded range). (E) Scatter plot showing correlation between binding Ab titres and pseudo-neutralising responses. Pooled data from WT and all SARS-CoV-2 variants analysed. Correlation coefficient (rho) was determined using Spearman rank correlation. Red dashed line represents locally weighted scatterplot smoothing (LOWESS). (F) Representative FACS-plot of AIM assay gated on CD4+ T cells. (G) Bar plots showing median stimulation index of CD4+ T cells against Wu-Hu.1 and XBB.1.5 calculated as a ratio between AIM + cells in peptide-stimulated wells and controls (median with 95% CI). For T cell analysis, HC (n = 23), PID (n = 27), HIV (n = 27), HSCT (n = 50), SOT (n = 22), and CLL (n = 31). The star annotation (∗) indicates statistical significance at a p-value threshold of 0.05 (or ∗∗ for p < 0.01, ∗∗∗ for p < 0.001, ∗∗∗∗ for p < 0.0001).