2.1 Study design and population

This is a prospective, longitudinal, and observational study conducted by Vall d'Hebron University Hospital Campus in Barcelona, Spain. The study cohort included patients diagnosed with hematologic malignancies or who underwent an allogenic HSCT (median time from transplant to vaccination 29.6 months, range 3.7–70.3 months) who received the mRNA-1273 SARS-CoV-2 vaccine. The initial scheme included two vaccine doses administered 28 days apart during the months of March and April 2021 and a posterior third vaccine dose between September 2021 and December 2021 (median of 26 weeks [interquartile range (IQR) 26–31] apart from the second vaccine dose).

Patients who suffered COVID-19 prior or during the study and positive anti-S above the selected threshold or developed T-cell response were also included in the analysis.

Blood draws were obtained at four points: (1) Baseline, immediately before the first vaccine dose (March 2021); (2) post-second dose (post-D2), after a median time of 26 days (IQR 22–28) (April/May 2021); (3) pre-third dose (pre-D3), at a median time of 23 weeks (IQR 22–23) after the second vaccine dose (September/October 2021); and (4) post-third dose (post-D3), at a median time of 11 weeks (IQR 8–15) after the third dose (December 2021/February 2022) (Figure 1).

To determine the immunological condition, a blood count test was performed, and total serum immunoglobulin levels were evaluated at baseline and pre-D3. In addition, at the latter time point, T, B, and natural killer cell subpopulations were also assessed.

In addition, to identify individuals with previous or recent SARS-CoV-2 asymptomatic infection, specific antibodies against the nucleocapsid (N) were measured at every time point.

This study was approved by the Institutional Clinical Research Ethics Committee of our center (Study number: 5818) and all patients provided written informed consent in accordance with the Helsinki declaration.

2.2 Assessment of humoral immune response

To evaluate the humoral immune response to the vaccine, at each time point two commercial chemiluminescence immunoassays (CLIA) were used: (1) LIAISON® SARS-CoV-2 TrimericS immunoglobulin G (IgG) (DiaSorin, Stillwater, MN, USA) performed on the LIAISON® XL Analyzer (DiaSorin, Italy) for the determination of IgG antibodies against spike glycoprotein of SARS-CoV-2 (range < 4.81 to >2080 BAU/mL), and (2) Elecsys® Anti-SARS-CoV-2 (Roche Diagnostics, Mannheim, Germany) performed on the Cobas® 8800 system (Roche Diagnostics, Switzerland) for the determination of total antibodies (including IgG, immunoglobulin M [IgM], and immunoglobulin A [IgA]) against nucleocapsid SARS-CoV-2 proteins (cutoff = 1.0 index).

A positive humoral response or seroconversion was defined when an anti-S IgG titer was ≥260 BAU/mL based on the agreement to promote an international common measure unit,^{23, 24} the correlation of anti-S IgG level ≥ 260 BAU/mL with 80% vaccine efficacy against symptomatic COVID-19,,²⁵ and also on the clinical impact of this threshold in the Spanish health care system that enables prescription of anti-SARS-CoV-2 monoclonal antibodies pre- and post-exposure to patients considered as nonresponders below this antibodies level.

2.3 Assessment of cellular immune response

SARS-CoV-2-specific T-cell response was evaluated at each time point by the whole blood interferon-gamma-release-immuno-assay (IGRA) technology using QuantiFERON® SARS-CoV-2 RUO tubes from Qiagen® (Hilden, Germany), consisting of two tubes coated with a combination of SARS-CoV-2 spike antigens (S1, S2, RBD), a mitogen tube that serves as a positive control and a Nil tube as a negative control.²⁶ This test was conducted following the manufacturer's instructions: venous blood samples were collected directly into the four QuantiFERON® SARS-CoV-2 RUO tubes, incubated at 37°C for 16–24 h and centrifuged to separate plasma. Interferon-gamma (IFN-γ) (IU/mL) was measured by CLIA using LIAISON® QuantiFERON-®TB Gold Plus assay (DiaSorin, Saluggia, VC, Italy) with the LIAISON® XL Analyzer (DiaSorin, Italy). Experimentally established cut-off values (Ag1 = 0.06175²⁷ and Ag2 = 0.05135 [unpublished data]) were used for the interpretation of the results.

2.4 Assessment of vaccine effectiveness

The incidence of COVID-19 was evaluated during a median follow-up of 13.8 months (95% confidence interval [CI] 13.7–14 months) since the administration of the first vaccine dose. Patients with compatible COVID-19 symptomatology underwent a screening with a SARS-CoV-2 RT-PCR or an antigen test. Viral variants were genetically characterized in the RT-PCR confirmed cases and clinical data about the severity of COVID-19 (World Health Organization [WHO] classification), rate of hospitalization, ICU admission, need for supplemental oxygen or COVID-19 specific therapy (steroids, antivirals, IL-6 inhibitors, anti-SARS-CoV-2 monoclonal antibodies) were collected. We considered an asymptomatic infection the observation of an anti-N seroconversion between two consecutive serological tests.

2.5 Statistical analysis

A descriptive analysis of all included variables in the study was performed. Continuous variables were expressed as median and IQR and categorical variables were described as absolute values and percentages. Univariate logistic regression models were carried out to estimate the association between baseline and sixth-month clinical factors, and between the early, persistent, and post-third vaccine dose cellular and humoral immunization of the mRNA-1273 vaccine. Odds ratios (ORs) with 95% CIs were reported. No data imputation was performed and the data analyses were carried out using R statistical software version 3.6.2. For original data, please contact the corresponding author.

3.1 Patient characteristics

A total of 270 patients were included in this study. Baseline and pre-third vaccine dose demographic and clinical characteristics from patients are summarized in Table S1.

During the time, post-D2 and pre-D3, 26 patients (9.6%) changed their hematologic disease status (11 patients had a progression, 10 patients in stable disease achieved a complete response, and 5 treatment naïve patients initiated treatment and attained a partial response). Also, 37 patients changed their treatment status (13.7%): 19 patients under active therapy changed their initial treatment to another therapy, 12 patients completed their therapy prior to the third dose and 6 patients off therapy/treatment naïve initiated active therapy during the observation study period.

3.2 Persistence of humoral and cellular immunogenicity 6 months after two vaccine doses

Two vaccine doses elicited an early humoral response rate of 68.5% (185/270) with a median anti-S of 1577 BAU/mL (IQR 96 to >2080) at 4 weeks after vaccination, including the seropositive patients with prior COVID-19 infection (n = 12) and the change in the threshold of humoral responders (anti-S IgG titer ≥260 BAU/mL). Patients older than 65 years and patients diagnosed with CLL presented a seroconversion rate of 66.9% and 65.4%, respectively. Nevertheless, their median anti-S title was significantly lower (1352 BAU/mL and 650 BAU/mL, respectively, p < .05) post-D2.

After almost 6 months (median of 23 weeks) post-D2, a decline in the seroconversion rate from 68.5% to 59.3% (147/248) and a reduction in the median anti-S antibodies titers from 1577 BAU/mL to 456 BAU/mL (IQR 33–1433) were observed. Of note, 31 out of 185 (16.8%) initially seropositive patients became seronegative. The factors associated with losing the humoral response were: >65 years of age (reduction from 66.9% to 50.9% of the seropositivity rate; p = .04), CLL as an underlying disease (reduction from 65.4% to 44.7% of the seropositivity rate; p = .04), active therapy (reduction from 53.5% to 41.4% of the seropositivity rate, p = .03) and treatment with BTK inhibitors (reduction from 56.3% to 25.9% of the seropositivity rate, p < .001). From the quantitative standpoint, the anti-S waning was observed in the whole cohort despite type of malignancy, status or type of treatment (Figure 2A, C, E).

Other factors associated with a negative humoral immunogenicity at 6 months post-D2 are described in Figure 3A. By contrast, only four patients initially negative became seropositive at this time point: two of them due to an asymptomatic SARS-CoV-2 infection with corroborated anti-N positivization, one probably caused by traces of anti-S antibodies present in transfused plasma or platelets units, and the last one due to a probable delayed humoral response to vaccination since the initial evaluation was early at Day 12 post-D2.

Regarding the cellular response, including the patients with prior COVID-19 infection and positive interferon-γ producing SARS-CoV-2-reactive T cells (n = 25), a total of 227 out of 270 (84.1%) patients had an early cellular immunity 4 weeks after the second dose of the mRNA-1273 vaccine with a median Ag1 of 0.4 IU/mL (IQR 0.09–1.28). Cellular response was more stable than the humoral immunogenicity maintaining a positivity rate of 84.4% (205/243) 6 months post-D2 with a median Ag1 of 0.52 IU/mL (IQR 0.1–1.77) (Figure 2B, D, F). During this time, only 17 patients (7.49%) lost their initial positive cellular response. Factors associated with a negative cellular response at this time point were: active therapy; treatment with immunomodulatory drugs (IMIDs), other target therapies and immunosuppressive agents; low IgA; and low lymphocyte, CD4⁺, CD8⁺, CD19⁺, and CD56⁺ cell counts (Figure 3B).

Moreover, 16 patients (7.8%) initially categorized as negative responders became positive at this time point without clinical or serological evidence of SARS-CoV-2 infection, all of them with negative anti-N antibodies. This subgroup of late cellular responders was composed predominantly by patients older than 65 years of age (63%) and with lymphoproliferative malignancies (5 CLL and 5 lymphoma patients).

3.3 Humoral and cellular immunogenicity after a third vaccine dose

A third vaccine dose was successful to improve the seroconversion rate from 59.3% to 75.2% (176/234) and increase the median anti-S titer from 456 BAU/mL to >2080 BAU/mL (IQR 289 to >2080). Overall, the third vaccine dose achieved a seroconversion of 42.7% (41/96) of patients without humoral response pre-D3. Regarding the subgroup of patients that did not have early humoral response post-D2, the third dose attained a seroconversion in 27.1% (19/70) of them. These patients were diagnosed mainly with CLL (42% [8/19], most of them (6/8) under BTK treatment) or lymphoma (42% [8/19] of whom 4 had finished their anti-CD20 therapy close to the initial vaccination scheme but it was ≥6 months prior to the third vaccine dose). In addition, 84.6% (22/26) of patients who had lost the response at the 6th month post-D2 regained the seropositivity. The seroconversion after the third dose was mostly attributable to vaccination, of the 16 seronegative patients post-D2 who developed SARS-CoV-2 infection only 5 had detectable anti-S antibodies and 3 anti-N positive antibodies post-D3.

Factors associated with a lack of humoral post-D3 were: lymphoma as underlying disease, active treatment, therapy with anti-CD20 antibodies, lymphopenia, hypogammaglobulinemia and, low CD3⁺ and CD19⁺ cell count (Figure 4A), being the most important factor the anti-CD20 therapy. Thus, 33 out of 58 patients (59.9%) with absence of seroconversion were receiving active treatment with anti-CD20 antibodies, and only 2 out of 35 patients (5.7%) under active treatment with these antibodies achieved a humoral response.

The overall cellular immunogenicity rate remained stable post-D3 with an 84.2% of responders (176/209). Overall, 14 out of 168 (8.3%) patients previously positive lost their response post-D3, while 11 out of 30 (36.7%) previous nonresponders achieved cellular response. The cellular responders post-D3 were composed of 4 out of 17 patients (23.5%) primarily negative post-D2; 2 were lymphoma patients who finished their chemoimmunotherapy ≥6 months prior to this third vaccine dose and improved their lymphocyte count; however, the other 2 patients were allogeneic stem cell transplant (allo-SCT) who developed COVID-19 and also positivized anti-N antibodies. The third dose also regained T-cell response in 7 out of 13 (53.8%) patients who lost it at 6 months post-D2, 4 of them were MM patients and 3 CLL patients, most of them under active therapy.

Factors associated with a negative cellular immune response post-D3 were: absence of response to therapy with progression or stable hematologic disease, treatment with GvHD immunosuppressive therapy, treatment with targeted therapies (mainly venetoclax and daratumumab), lymphopenia and low CD3⁺, CD4⁺, and CD56₊ cell count (Figure 4B).

Importantly, post-D3 only 7.2% (15/209) of the patients lacked both humoral and cellular vaccine-induced immunogenicity, most of them being lymphoma patients under active treatment (8/15), mainly with anti-CD20 antibodies (6/8).

3.4 Vaccine effectiveness against SARS-CoV-2 infection and severe COVID-19

During a median observation period of 13.8 months from March 2021 to May 2022 (since the administration of the first vaccine dose until 5 months post-D3), 67 out of 270 patients (24.8%) developed breakthrough SARS-CoV-2 infection despite vaccination, 19.4% (13/67) of them being asymptomatic. Twenty patients (29.9%) were infected post-D2 (May 2021–December 2021) and 47 patients (70.1%) post-D3 (October 2021–May 2022). Among the symptomatic cases (n = 54), 55.6% were diagnosed through rapid antigen test and 44.4% (24/54) through RT-PCR; regarding the latter, viral sequencing from 18 cases disclosed that the viral variant was delta in 5 (28%) and omicron in 13 (72%).

Age > 65 years (OR 0.56 (95% CI 0.31–0.99); p = .05) and the development of a positive serological response post-D2 (OR 0.50 (95% CI 0.28–0.89); p = .02) resulted associated with less risk to develop breakthrough SARS-CoV-2 infection.

Most of the infections were mild or moderate according to the WHO severity of illness scale (71.6%; 48 out of 67). The hospitalization rate in all infected patients was 13.4% (9/67), 9% (6/67) of patients needed supplemental oxygen and the ICU admission rate was 4.5% (3/67). The COVID-19-specific therapies administered were remdesivir (16.4%), steroids (9%), tocilizumab (3%), convalescent plasma (3%), and anti-SARS-CoV-2 monoclonal antibodies (1.5%).

Six out of 67 (9%) patients suffered a severe/critical infection, including 4 patients with lymphoma, 1 patient with CLL, and 1 allo-SCT patient; 5 of them were under active therapy (3 with anti-CD20 antibodies, 1 with BTK inhibitor, and 1 with PI3K inhibitor) and in therapeutic response. Three cases were infected by omicron variant, 1 by delta variant, and 2 could not be sequenced. It is worth mentioning that none of the severe cases had developed humoral vaccine response, while 3 out of 6 also did not develop cellular immunogenicity. Moreover, patients with a severe infection had a weaker cellular response measured by a lower median IFN-γ production (0.19 IU/mL [IQR 0.02–0.35] versus 0.35 IU/mL [IQR 0.10–1.24]) and lower median anti-S titer (<4.81 BAU/mL [IQR < 4.81–30.3]) versus 788 BAU/mL (IQR 32.11 to >2080) in comparison with asymptomatic/mild cases at the last assessment previous infection. Two out of the 6 patients who developed severe disease died due to COVID-19.

Finally, 6 out of 39 (15.4%) patients developed a prolonged virus shedding (positive RT-PCR >30 days), 5 of those patients had a negative humoral response and 4 had negative cellular immunogenicity. Three out of those 6 cases (50%) suffered a severe COVID-19.