Abstract
Objective To determine how serologic responses to coronavirus disease 2019 (COVID-19) vaccination and infection in immune-mediated inflammatory disease (IMID) are affected by time since last vaccination and other factors.
Methods Post–COVID-19 vaccination, data, and dried blood spots or sera were collected from adults with rheumatoid arthritis, inflammatory bowel disease, systemic lupus erythematosus, ankylosing spondylitis and spondylarthritis, and psoriasis and psoriatic arthritis. The first sample was collected at enrollment, then at 2 to 4 weeks and 3, 6, and 12 months after the latest vaccine dose. Multivariate generalized estimating equation regressions (including medications, demographics, and vaccination history) evaluated serologic response, based on log-transformed anti–receptor-binding domain (RBD) IgG titers; we also measured antinucleocapsid (anti-N) IgG.
Results Positive associations for log-transformed anti-RBD titers were seen with female sex, number of doses, and self-reported COVID-19 infections in 2021 to 2023. Negative associations were seen with prednisone, anti–tumor necrosis factor agents, and rituximab. Over the 2021-2023 period, most (94%) of anti-N positivity was associated with a self-reported infection in the 3 months prior to testing. From March 2021 to February 2022, anti-N positivity was present in 5% to 15% of samples and was highest in the post-Omicron era, with antinucleocapsid positivity trending to 30% to 35% or higher as of March 2023. Anti-N positivity in IMID remained lower than Canada’s general population seroprevalence (> 50% in 2022 and > 75% in 2023). Time since last vaccination was negatively associated with log-transformed anti-RBD titers, particularly after 210 days.
Conclusion Ours is the first pan-Canadian IMID assessment of how vaccine history and other factors affect serologic COVID-19 vaccine responses. These findings may help individuals personalize vaccination decisions, including consideration of additional vaccination when > 6 months has elapsed since last COVID-19 vaccination/infection.
- autoimmune diseases
- inflammatory bowel disease
- psoriatic arthritis
- rheumatic diseases
- systemic lupus erythematosus
- vaccination
SARS-CoV-2 precipitated a global crisis in 2020; unfortunately, this virus has not been eliminated, but is now endemic. In 2022-2023, 4 times as many people in the United States were hospitalized for coronavirus disease 2019 (COVID-19) compared to influenza, and COVID-19 mortality risk in this setting remains twice that of influenza.1 Thus, we continue to require information regarding serologic responses and breakthrough infections after COVID-19 vaccination.
Currently, individuals with immune-mediated inflammatory disease (IMID) may be unsure about the value of COVID-19 vaccination beyond the primary series. This is particularly concerning since immunosuppressant therapy may put individuals at higher risk for SARS-CoV-2 transmission.
Vaccine hesitancy in general is a serious issue, particularly for people with IMID.2 As the necessity for multiple COVID-19 vaccine doses became clear, social “vaccine fatigue” has caused many individuals to decline additional COVID-19 boosters after the primary series.3 Since COVID-19 infection is a potentially fatal comorbidity that can be mitigated by vaccination, it is vital that individuals with IMID have access to relevant information that will help them decide when to get their next COVID-19 vaccination.
To address this need, we evaluated how COVID-19 vaccination history (including time since last vaccine) and other factors influence serological response to COVID-19 vaccination and infection.
METHODS
The Canadian government’s COVID-19 Immunity Task Force4 funded our study of COVID-19 vaccination responses in IMID. Participants were recruited from Vancouver, Calgary, Winnipeg, Montreal, Quebec City, Sherbrooke, Toronto, and Hamilton. Baseline and follow-up questionnaires (paper or electronic) and dried blood spots (DBS; or sera in Calgary and Winnipeg) were collected before and after each COVID-19 vaccination dose following enrollment. Initial recruitment began early in 2021, a few months after Canada began vaccinating against SARS-CoV-2 (mostly using mRNA vaccines). Recruitment at most centers initially targeted assessments at predose for vaccines 1 and 2 (for mRNA formulations, which were anticipated to be the vast majority) and at 2-4 weeks, then 3 and 6 months later. Soon after beginning recruitment, it became clear that patients would receive third and subsequent vaccinations, so we amended our protocol to enroll individuals at any point up to 6 months after their last vaccination. To reflect real-world experience, we allowed any consenting adult with a clinical IMID diagnosis (by a physician) to be enrolled regardless of therapy. Individuals had to have been vaccinated against SARS-CoV-2 within 6 months of enrollment or be planning further vaccination. The study was approved by the ethics board of the research institute of the McGill University Health Center (MP-37-2022-7763) and all participating centers.
Participants provided baseline information on past COVID-19 infection, COVID-19 vaccinations (including dates and type), and clinical history (type of IMID, medications). Our protocol included collection of data and DBS/sera at enrollment, 2-4 weeks, and 3, 6, and 12 months after COVID-19 vaccination.
Post enrollment, participants were asked to contact the research team if they received additional vaccine doses or if they developed a COVID-19 infection. They were then asked to provide updated data and DBS/sera 2-4 weeks post vaccine or infection. Participants were also contacted at 3, 6, and 12 months to update information on medications and confirm whether they had experienced additional COVID-19 infections or vaccinations. If participants had a new vaccination during follow-up, they reverted to collecting biospecimens for serological sampling at 2-4 weeks and 3, 6, and 12 months post vaccination, up to end of sampling (July 15, 2023).
DBS/sera collected by participants at home were mailed (in prepaid envelopes) back to each participating site, and sent in batches to the Gingras lab for antibody testing as described previously; in brief, samples processed in the Gingras lab were tested with automated ELISAs for antibodies (IgG) to the spike trimer (SmT1; anti-S), the antireceptor-binding domain (anti-RBD), and antinucleocapsid (anti-N) using standardized assays.5 Exceptions in our study were individuals from Winnipeg and Calgary, who collected sera at their center that underwent local assays as previously described6 (in Calgary, sera were assayed for anti-RBD and anti-N but not anti-S antibodies).7,8
We performed descriptive analyses of anti-RBD/anti-S serologic responses and used multivariate generalized estimating equation (GEE)9 regressions (accounting for repeated measures) to evaluate log-transformed anti-RBD titers (as the Calgary center did not assay for anti-S). We then exponentiated the β coefficients for each covariate in the univariate and multivariate models to generate parameter estimates for the effect of each factor on the outcome, with 95% CIs. The Shapiro-Wilk normality test indicated that the data approximated the normal distribution after log-transformation of anti-RBD titers.
Our GEE models adjusted for baseline demographics (age at enrollment as a continuous variable, sex, White vs all other races/ethnicities), type of IMID, recruiting centers, and current relevant medication exposures (at recruitment). Medications of interest included prednisone (which we treated as categorical), biologics (with categories for anti–tumor necrosis factor [TNF], rituximab [RTX], and other agents [eg, ustekinumab, vedolizumab, belimumab]), and other immunomodulators (with categories for methotrexate [MTX], Janus kinase inhibitors [JAKi], and other immunomodulators [eg, azathioprine]).
Time-varying variables in our multivariate models included vaccine type, time between specimen collection and last vaccination, and whether patients self-reported a COVID-19 infection (for which the patients had a positive clinical test result; some patients could report more than 1 infection over time, and our model indicated if they had reported > 1 infection per year). Vaccination type was assigned categorically according to whether a subject had received exclusively either 1 of 2 available monovalent mRNA formulations (BNT-162b2 or mRNA1273) vs any past combination of the 2 monovalent mRNA formulations vs any bivalent mRNA (available in Canada from the fall of 2022 onward), the reference being samples collected only after vaccination with non-mRNA formulations (eg, AZD1222 [ChAdOx1] or Ad26.COV2.S).
We also described anti-N IgG seropositivity, which was initially considered as a marker of recent infection10 in the literature, and we compared anti-N IgG seropositivity to self-reported infection (in the 3 months prior to sample). We calculated percent of samples demonstrating anti-N IgG seropositivity within any given month, from 2021 onward. Anti-N was not part of our multivariate modeling of anti-RBD titers.
RESULTS
The characteristics of 1823 participants according to their absence or presence of any anti-RBD or anti-S antibodies post-vaccination are shown in Table 1. To summarize, approximately two-thirds (64.7%) were female, the mean (SD) age was 53.2 (15.3) years, and the majority (88.4%) were White. The most common IMID was IBD, followed by rheumatoid arthritis (RA). Approximately 11% reported COVID-19 infection in the 3 months before providing a sample; comparing self-reported infection across years, the highest number was in 2022 (49.1% of participants reported infection that year). At baseline (cohort enrollment), just under 25% of participants were on prednisone, about 61% were on a biologic, the most common being anti-TNF agents (33.6% of participants at enrollment). The most frequent nonbiologic drug at enrollment was MTX (29.6% of the sample). The most common type of vaccine received was BNT-162b2 monovalent (59.4%). Table 1 shows that 210 people were RBD/SmT1 negative following their last vaccination; two-thirds (140/210) of these had only provided 1 sample and the remainder provided 2 or 3 samples. Approximately 40% of these 210 (n = 81) with negative samples were within 3 months of their last vaccination, whereas about 35% of these 210 (n = 74) were within 3-6 months of their last vaccination. The remainder were > 6 months since their last vaccination.
Characteristics of participants according to whether they had any antibodies to SARS-CoV-2 spike trimer (SmT1) and/or RBD across time.a
Table 2 shows crude and adjusted results of the GEE multivariate model (with the outcome of log-transformed anti-RBD titer). The exponentiated β coefficient for each variable can be interpreted as representing a relative decrease in the outcome; for example, an exponentiated β coefficient of 0.90 suggests a 10% decrease relative to the reference and an exponentiated β coefficient of 1.10 suggests a 10% increase relative to the reference. Positive associations with immunogenicity (log-transformed anti-RBD titers) were seen with female sex, number of COVID-19 vaccine doses, and COVID-19 infections in 2021 and later. Time since last vaccination was negatively associated with log-transformed anti-RBD titers. In sensitivity analyses, the GEE model adjusted exponentiated coefficient was statistically lower after 210 days and beyond (exponentiated coefficient 0.88 [95% CI 0.80-0.95]).
Exponentiated coefficients and 95% CIs for the effects of demographics, clinical exposures, and vaccination history on anti-RBD serology.
Negative associations with immunogenicity (log-transformed anti-RBD titers) were seen with use of prednisone (particularly at doses of ≥ 20 mg), anti-TNF agents, and RTX. Of those reporting the use of prednisone, the mean dose was 21 mg (median 10, IQR 5-25 mg, minimum dose 0.5 mg, maximum dose 120 mg). Of 9 patients with systemic lupus erythematosus (SLE) receiving belimumab (included in the “other biologic” category), all were positive for SARS-CoV-2 antibodies.
The percent of samples testing positive for anti-N over time is shown cross-sectionally by 2-month intervals in the Figure. From March 2021 to January 2022, anti-N positivity was present in 5% to 15% of samples; in February to September 2022, anti-N positivity trended higher at 20% to 27%. Overall, anti-N positivity was lowest in the summer of 2021 (about 5%) and highest in the post-Omicron era, which began in Canada in November 2021, and was associated with anti-N positivity trending to 30% to 35% or higher as of March 2023. The vast majority (94%) of people with anti-N antibodies reported having had an infection in the 3 months before the sample.
Cross-sectional percent of antinucleocapsid (anti-N) positivity in samples over time.
DISCUSSION
In North America and other parts of the world, SARS-CoV-2 infection continues to be a common cause of hospitalization and, in absolute terms, causes more deaths than influenza. The topic of serologic response and breakthrough after COVID-19 vaccination remains very timely as individuals with IMID may be unsure about the value of additional COVID-19 vaccine doses. This is particularly important in IMID, where immunosuppressant therapy puts individuals at higher risk for SARS-CoV-2 transmission.
In general, vaccine hesitancy and low vaccine coverage in IMID have been highly relevant issues for years. Unfortunately, ensuring adequate vaccination against SARS-CoV-2 in IMID can be difficult, particularly when individuals decline additional COVID-19 vaccine doses after the primary series. If individuals with IMID have access to relevant information that will help them to decide when to get their next COVID-19 vaccine, they may be more open to receiving additional doses. We believe that our findings will help patients, clinicians, and other stakeholders make personalized vaccination decisions.
Regarding time since last COVID-19 vaccination, our sensitivity analyses indicated that antibody levels appear to significantly decrease when 210 days or more had passed since the last vaccination. The number of COVID-19 vaccine doses was positively associated with log-transformed anti-RBD titers. As seen in Table 2, fourth and fifth doses did positively affect immunogenicity. Our data suggest that these individuals should continue to consider additional doses when more than 6 months has elapsed since last vaccination or infection.
In our detailed multivariate analyses of this large, relatively unselected IMID sample, we found reduced responses in patients on certain immune drugs, which resonates with earlier studies showing relatively lower antibodies (and neutralizing ability) in patients with IMID receiving anti-TNF therapies.11 Public health recommendations12 consider 3 mRNA COVID-19 vaccinations as the primary series in immunocompromised individuals (as opposed to 2 mRNA vaccines, which were considered the primary series in nonimmunocompromised individuals in Canada)13 and further doses in immunocompromised individuals.14
In our study, individuals exposed to anti-TNF agents, RTX,15 and prednisone (> 20 mg) all demonstrated less immunogenicity than those not exposed to these drugs, and this was evident even adjusting for demographics and other clinical factors, such as concomitant MTX and other immunosuppressives. Other studies have also suggested that exposure to these drugs contribute to reduced serologic responses to COVID-19 vaccination.11 Specific effects of anti-TNF agents in decreasing serologic responses to COVID-19 vaccination have been shown in rheumatic diseases and in IBD, but to our knowledge, ours is the first Canadian multiprovince assessment of the modeling of concomitant drug effects, with adjustment for other vaccine-related and demographic and clinical factors. Of 9 patients with SLE on belimumab, all were positive for SARS-CoV-2 antibodies.
RTX in particular has been considered a key depressor of immune response to COVID-19 vaccination; many earlier studies had not focused solely on IMID, nor assessed responses beyond the first year of vaccination.16 Unfortunately, the data did not allow us to examine timing of RTX administration with respect to vaccination, and how serologic results may have been affected in this regard. It should be noted that IMID drug exposures do tend to be somewhat disease-specific; for example, RTX is primarily used in RA and sometimes in SLE (but not in IBD, psoriatic arthritis [PsA], psoriasis [PsO], or spondyloarthritis [SpA]), whereas anti-TNF agents are often used in both RA and IBD (as well as PsA/PsO or SpA) but not in SLE. This was a benefit of combining data across a number of different IMIDs.
It must be acknowledged that information on past COVID-19 infection, COVID-19 vaccinations, and clinical history were self-reported and could, due to recall error, contribute to misclassification of some of these exposure variables. In our assessment, this would likely be nondifferential by outcome, which could have contributed to biases toward the null for estimates of the effects of some of these variables. However, most participants were enrolled relatively early on in the pandemic, when maintaining proof of vaccinations was important for employment and social activities. Another potential limitation was that only symptomatic infections were recorded. Asymptomatic or subclinical COVID-19 infection would be difficult to document. This could have led to nondifferential misclassification of that exposure.
We had relatively few individuals receiving JAKi, which limited our ability to establish specific effects of these agents. We did not note a lower serologic response in individuals exposed to MTX in our adjusted model. We are not able to comment on whether stopping MTX (or any other agent) before vaccination would improve response.
The positive association between female sex and COVID-19 immunogenicity is interesting. It is well established as a general principle that females in the general population develop greater antibody responses following vaccinations of various types. This is possibly due to stronger and more rapid immune responses in female vs male individuals.17
Past COVID-19 infections were positively associated with log-transformed anti-RBD titers; this has also been shown in other studies.18 Public health measures in 2022-2023 deferred vaccination some months after a COVID-19 infection.19 The debate has endured as to the quality of response post infection (vs post vaccination) and to what extent it matters where in the course of vaccination (before, during, or after the primary series) the infection occurs, as well as whether the infection occurred in the post-Omicron era or not.20
Breakthrough infections, as noted by self-report and possibly by anti-N serology, were considerable and accumulated over time—a trend that has also been seen in Canada’s general population—although anti-N positivity was lower in our IMID sample compared to Canada’s general population (estimated at > 50% in July 2022 and > 75% in January 2023).21 The demographics and methods of sampling were different for Canada’s general population estimates vs ours, which may explain some of the differences. It is also probable that people with IMID practiced COVID-19 precautions more strictly and for longer periods than the general population.
A strength of our study is that it provides a detailed assessment of drugs, vaccine history, and other factors on serologic COVID-19 vaccine response in a large, pan-Canadian IMID sample. Individuals were enrolled from multiple centers across Canada, thus hopefully ensuring a fairly representative, real-world sample. One potential limitation is that, although we did collect detailed drug exposure information, in the current analyses it is difficult to specifically analyze issues such as whether a patient had held a drug (eg, MTX) in the week before or after vaccination. Also, we did not assess complex interactions, either between drugs, between drugs and IMID type, or between vaccination type and COVID-19 infection.
In summary, the topic of serologic response and breakthrough infections after COVID-19 vaccination remains very timely. Ensuring adequate vaccination against SARS-CoV-2 in IMID can be difficult. If individuals have access to relevant information that will help them decide when to get their next COVID-19 vaccine, they may be more open to receiving additional doses. We believe that our findings will help patients, clinicians, and other stakeholders make personalized vaccination decisions, including consideration of additional doses when more than 6 months has elapsed since last COVID vaccination or infection.
ACKNOWLEDGMENT
This project was supported by funding from the Public Health Agency of Canada, through the Vaccine Surveillance Reference group and the COVID-19 Immunity Task Force. The views expressed here do not necessarily represent the views of the Public Health Agency of Canada. We acknowledge the time, effort, and dedication of all members of the various research teams as well as patients recruited for the study.
Footnotes
This project was supported by funding from the Public Health Agency of Canada, through the Vaccine Surveillance Reference group and the COVID-19 Immunity Task Force.
CNB is supported by the Bingham Chair in Gastroenterology and has served on advisory boards for AbbVie, Amgen, BMS, Eli Lilly, Ferring, JAMP, Janssen, Pendopharm, Sandoz, Takeda, and Pfizer; been a consultant for Mylan and Takeda; served on speakers’ panels for AbbVie, Janssen, Pfizer, and Takeda; received educational grants from AbbVie, Amgen, BMS, Eli Lilly, Ferring, Pfizer, Takeda, and Janssen, and received research funding from AbbVie, Amgen, Pfizer, Sandoz, and Takeda (none relevant for this work). GB has received honoraria (none relevant to this work) for speaking or consultancy from AbbVie, BMS, Lilly, Novartis, Pfizer, Samsung BioEpis, Viatris; multicentric research grants (none relevant to this work) from Janssen and Pfizer; and unrestricted grant support (none relevant to this work) for local initiatives from BMS, Lilly and Pfizer. DMEB is the Canada Research Chair in Aging & Immunity. VC has received research grants (none relevant to this work) from AbbVie, Amgen, and Eli Lilly, and has received honoraria (none relevant to this work) for advisory board member roles from AbbVie, Amgen, BMS, Eli Lilly, Janssen, Novartis, Pfizer, and UCB. VC’s spouse is an employee of AstraZeneca. ACG has received research funds from a research contract with Providence Therapeutics Holdings, Inc., for other projects; participated in the COVID-19 Immunity Task Force (CITF) Immune Science and Testing working party; chaired the CIHR Institute of Genetics Advisory Board; and currently chairs the SAB of the National Research Council of Canada Human Health Therapeutics Board. ACG is also the Canada Research Chair, Tier 1, in Functional Proteomics. CAH received research grants for multicentric research from AstraZeneca, and research grants from Pfizer. GGK has received honoraria (none relevant to this work) for speaking or consultancy from AbbVie, Amgen, Janssen, Pfizer, Sandoz, and Pendophram; grants (none relevant to this work) for research from Ferring and for educational activities (none relevant to this work) from AbbVie, BMS, Ferring, Fresenius-Kabi, Janssen, Pfizer, and Takeda. DPR is the volunteer Vice President of the Canadian Arthritis Patient Alliance, a not-for-profit organization run by and for patients that receives most of its funding in the form of independent grants from pharmaceutical companies. THW is the Canada Research Chair in Antiviral Immunity.
- Accepted for publication March 13, 2024.
- Copyright © 2024 by the Journal of Rheumatology
This is an Open Access article, which permits use, distribution, and reproduction, without modification, provided the original article is correctly cited and is not used for commercial purposes.
