Safety and immunogenicity of one versus two doses of the COVID-19 vaccine BNT162b2 for patients with cancer: interim analysis of a prospective observational study

Abstract

Background: The efficacy and safety profiles of vaccines against SARS-CoV-2 in patients with cancer is unknown. We aimed to assess the safety and immunogenicity of the BNT162b2 (Pfizer-BioNTech) vaccine in patients with cancer.

Methods: For this prospective observational study, we recruited patients with cancer and healthy controls (mostly health-care workers) from three London hospitals between Dec 8, 2020, and Feb 18, 2021. Participants who were vaccinated between Dec 8 and Dec 29, 2020, received two 30 μg doses of BNT162b2 administered intramuscularly 21 days apart; patients vaccinated after this date received only one 30 μg dose with a planned follow-up boost at 12 weeks. Blood samples were taken before vaccination and at 3 weeks and 5 weeks after the first vaccination. Where possible, serial nasopharyngeal real-time RT-PCR (rRT-PCR) swab tests were done every 10 days or in cases of symptomatic COVID-19. The coprimary endpoints were seroconversion to SARS-CoV-2 spike (S) protein in patients with cancer following the first vaccination with the BNT162b2 vaccine and the effect of vaccine boosting after 21 days on seroconversion. All participants with available data were included in the safety and immunogenicity analyses. Ongoing follow-up is underway for further blood sampling after the delayed (12-week) vaccine boost. This study is registered with the NHS Health Research Authority and Health and Care Research Wales (REC ID 20/HRA/2031).

Findings: 151 patients with cancer (95 patients with solid cancer and 56 patients with haematological cancer) and 54 healthy controls were enrolled. For this interim data analysis of the safety and immunogenicity of vaccinated patients with cancer, samples and data obtained up to March 19, 2021, were analysed. After exclusion of 17 patients who had been exposed to SARS-CoV-2 (detected by either antibody seroconversion or a positive rRT-PCR COVID-19 swab test) from the immunogenicity analysis, the proportion of positive anti-S IgG titres at approximately 21 days following a single vaccine inoculum across the three cohorts were 32 (94%; 95% CI 81-98) of 34 healthy controls; 21 (38%; 26-51) of 56 patients with solid cancer, and eight (18%; 10-32) of 44 patients with haematological cancer. 16 healthy controls, 25 patients with solid cancer, and six patients with haematological cancer received a second dose on day 21. Of the patients with available blood samples 2 weeks following a 21-day vaccine boost, and excluding 17 participants with evidence of previous natural SARS-CoV-2 exposure, 18 (95%; 95% CI 75-99) of 19 patients with solid cancer, 12 (100%; 76-100) of 12 healthy controls, and three (60%; 23-88) of five patients with haematological cancers were seropositive, compared with ten (30%; 17-47) of 33, 18 (86%; 65-95) of 21, and four (11%; 4-25) of 36, respectively, who did not receive a boost. The vaccine was well tolerated; no toxicities were reported in 75 (54%) of 140 patients with cancer following the first dose of BNT162b2, and in 22 (71%) of 31 patients with cancer following the second dose. Similarly, no toxicities were reported in 15 (38%) of 40 healthy controls after the first dose and in five (31%) of 16 after the second dose. Injection-site pain within 7 days following the first dose was the most commonly reported local reaction (23 [35%] of 65 patients with cancer; 12 [48%] of 25 healthy controls). No vaccine-related deaths were reported.

Interpretation: In patients with cancer, one dose of the BNT162b2 vaccine yields poor efficacy. Immunogenicity increased significantly in patients with solid cancer within 2 weeks of a vaccine boost at day 21 after the first dose. These data support prioritisation of patients with cancer for an early (day 21) second dose of the BNT162b2 vaccine.

Funding: King's College London, Cancer Research UK, Wellcome Trust, Rosetrees Trust, and Francis Crick Institute.

Figure 1

Serological response to COVID-19 vaccine BNT162b2 (A) Spike-specific IgG titres (EC₅₀) in plasma samples at 3 weeks after the vaccine in serological responders: healthy controls (n=32), patients with solid cancers (n=21), and patients with haematological cancer (n=8). The horizontal line represents the threshold of specific response. Short bars represent the median values of responder values only. Sample comparisons tested by Kruskal-Wallis with Dunn's post-hoc test on responder values only; no significant differences. (B) Association of age with serological response (spike-specific IgG ELISA) at 3 weeks after the vaccine (Spearman's correlation) in healthy controls (n=34; r=–0·1, p=0·58), patients with solid cancers (n=56; r=–0·12, p=0·47), and patients with haematological cancers (n=44; r=–0·11, p=0·66). The horizontal line represents the threshold of specific response. Dashed lines represent regression lines. Shading represents 95% CIs. (C) Neutralisation titres against wild-type SARS-CoV-2 (upper panel) and the B.1.1.7 variant of concern (lower panel) in plasma samples at 3 weeks after the vaccine in healthy controls (n=16), patients with solid cancer (n=14), and patients with haematological cancer (n=5). Short bars represent median values of responder values only. Sample comparisons tested by Kruskal-Wallis with Dunn's post-hoc test, corrected by Benjamini-Hochberg method; no significant differences. (D) Correlation between spike-specific IgG titres and neutralisation titres against wild-type SARS-CoV-2 samples (upper panels) at 3 weeks after the vaccine (Spearman's correlation) in healthy controls (n=16; r=0·18, p=0·00025) and patients with solid cancer (n=14; r=0·84, p=0·00028). Correlation between spike-specific IgG titres and neutralisation titres against B.1.1.7 SARS-CoV-2 samples (lower panels) at 3 weeks after the vaccine (Spearman's correlation) in healthy controls (n=16; r=0·55, p=0·030) and in patients with solid cancer (n=14; r=0·84, p=0·026). Dashed lines represent regression lines. Shading represents 95% CIs. EC₅₀=50% effective concentration. ID₅₀=inhibitory dilution at which 50% of viral particles are neutralised.

Figure 2

T-cell response to COVID-19 vaccine BNT162b2 (A) IFNγ⁺ and IL-2⁺ responses to stimulation with peptides from RBD, S2, and CEF/CEFT reported as number of spots per 10⁶ cells in PBMC samples at 3 weeks after the vaccine in healthy controls (n=17), patients with solid cancer (n=31), and patients with haematological cancer (n=18). Short bars represent median values for each group; the horizontal line represents the threshold of specific response. Kruskal-Wallis test with Dunn's post-hoc test, corrected by Benjamini-Hochberg method. p values shown where inter-group comparisons were significant: healthy control versus haematological cancer cohorts (IFNγ RBD, IFNγ S2, IL-2 RBD, IL-2 S2) and solid cancer versus haematological cancer cohorts (IL-2 RBD). (B) Relationship between serological response and T-cell response in healthy controls (n=17), patients with solid cancer (n=31), and patients with haematological cancer (n=18). The horizontal lines represent thresholds of S2-specific IFNγ T-cell responses and the vertical lines represent the thresholds of S-reactive serological responses. Square data points denote S2-specific IL-2 producers (ie, IL-2 threshold of >7 spots). No statistical test was done to assess the association between serological and T-cell responses because the plot serves to highlight the responder status of patients by threshold as a graphical representation. (C) Spearman's correlation between T-cell responses (fluorospot counts per 10⁶ PBMC) and serological responses as determined by ELISA and neutralisation assays across all study participants. The colour scale indicates Spearman's r value; all p values are less than 0·01. The circle sizes are proportional to the correlation coefficient. CEF/CEFT=cytomegalovirus, Epstein-Barr virus, influenza virus (and tetanus toxin) peptide pools. EC₅₀=50% effective concentration. ID₅₀=inhibitory dilution at which 50% of viral particles are neutralised. IFNγ=interferon-γ. IL-2=interleukin-2. PBMC=peripheral blood mononuclear cell. RBD=receptor binding domain. S=spike protein. S2=spike protein 2.

Figure 2

T-cell response to COVID-19 vaccine BNT162b2 (A) IFNγ⁺ and IL-2⁺ responses to stimulation with peptides from RBD, S2, and CEF/CEFT reported as number of spots per 10⁶ cells in PBMC samples at 3 weeks after the vaccine in healthy controls (n=17), patients with solid cancer (n=31), and patients with haematological cancer (n=18). Short bars represent median values for each group; the horizontal line represents the threshold of specific response. Kruskal-Wallis test with Dunn's post-hoc test, corrected by Benjamini-Hochberg method. p values shown where inter-group comparisons were significant: healthy control versus haematological cancer cohorts (IFNγ RBD, IFNγ S2, IL-2 RBD, IL-2 S2) and solid cancer versus haematological cancer cohorts (IL-2 RBD). (B) Relationship between serological response and T-cell response in healthy controls (n=17), patients with solid cancer (n=31), and patients with haematological cancer (n=18). The horizontal lines represent thresholds of S2-specific IFNγ T-cell responses and the vertical lines represent the thresholds of S-reactive serological responses. Square data points denote S2-specific IL-2 producers (ie, IL-2 threshold of >7 spots). No statistical test was done to assess the association between serological and T-cell responses because the plot serves to highlight the responder status of patients by threshold as a graphical representation. (C) Spearman's correlation between T-cell responses (fluorospot counts per 10⁶ PBMC) and serological responses as determined by ELISA and neutralisation assays across all study participants. The colour scale indicates Spearman's r value; all p values are less than 0·01. The circle sizes are proportional to the correlation coefficient. CEF/CEFT=cytomegalovirus, Epstein-Barr virus, influenza virus (and tetanus toxin) peptide pools. EC₅₀=50% effective concentration. ID₅₀=inhibitory dilution at which 50% of viral particles are neutralised. IFNγ=interferon-γ. IL-2=interleukin-2. PBMC=peripheral blood mononuclear cell. RBD=receptor binding domain. S=spike protein. S2=spike protein 2.

Figure 3

Comparison of single dose versus prime–boost with COVID-19 vaccine BNT162b2 (A) Spike-specific IgG titres in plasma samples at 3 and 5 weeks after the vaccine in individuals receiving a single vaccine dose (no boost) and in those receiving two doses (boost). Patients failing to achieve a serological response at any timepoint were excluded. Dashed lines represent eight non-responders with solid cancer at timepoint 2 who seroconverted following boost. (B) Neutralisation titres against wild-type SARS-CoV2 in plasma samples at 3 and 5 weeks after the vaccine in individuals receiving a single vaccine dose (no boost) and in individuals receiving two doses (boost). Patients failing to achieve a serological response at any timepoint were excluded. (C) Neutralisation titres against B.1.1.7 SARS-CoV-2 in plasma samples at 3 and 5 weeks after the vaccine in individuals receiving a single vaccine dose (no boost) and in individuals receiving two doses (boost). Patients failing to achieve a serological response at any timepoint are excluded. (D–F) Cytokine response to stimulation with peptides from RBD, S2, and CEF/CEFT reported as number of spots per 10⁶ PBMC at 3 and 5 weeks after the vaccine in individuals receiving a single vaccine dose (no boost) and in individuals receiving two doses (boost). All comparisons tested by paired Wilcoxon test, corrected by Benjamini-Hochberg method. CEF/CEFT=cytomegalovirus, Epstein-Barr virus, influenza virus (and tetanus toxin) peptide pools. EC₅₀=50% effective concentration. ID₅₀=inhibitory dilution at which 50% of viral particles are neutralised. IFNγ=interferon-γ. PBMC=peripheral blood mononuclear cell. RBD=receptor binding domain. S=spike protein. S2=spike protein 2.

Figure 4

Local and systemic effects reported within 30 days after injection of COVID-19 vaccine BNT162b2 in patients with cancer and healthy controls Data on local and systemic reactions were collected via telephone consultations with participants for 30 days after vaccination. (A) Proportion of participants reporting no toxicity or toxicity (local effects only vs systemic effect only vs both local and systemic effects) following the first dose and the second booster dose of BNT162b2 on day 21. (B) Breakdown of specific local and systemic side-effects in patients with cancer and healthy controls following the first dose. (C) Breakdown of specific local and systemic side-effects in patients with cancer and healthy controls following the second booster dose of BNT162b2 on day 21. Symptoms were assessed according to the following scale: grade 1 (mild; does not interfere with activity), grade 2 (moderate; interferes with activity), grade 3 (severe; prevents daily activity), and grade 4 (potentially life-threatening; emergency department visit or admission to hospital).