Duration of Postvaccination Neutralizing Antibodies to SARS‐CoV‐2 and Medication Effects: Results from the Safety and Immunogenicity of COVID‐19 Vaccination in Systemic Immune‐Mediated Inflammatory Diseases Cohort Study

Abstract

Objective

In the face of the ongoing circulation of SARS‐CoV‐2, the durability of neutralization post–COVID‐19 vaccination in immune‐mediated inflammatory disease (IMID) is a key issue, as are the effects of medications.

Methods

Adults (n = 112) with inflammatory bowel disease, psoriasis/psoriatic arthritis, rheumatoid arthritis, spondylarthritis, and systemic lupus were recruited from participating Canadian medical centers from 2021 to 2023. We focused on log‐transformed neutralization (lentivirus methods) as a continuous outcome, with separate models for wild‐type and Omicron strains BA.1 and BA.5.

Results

Compared with 30 to 120 days postvaccination, subsequent periods were associated with greater neutralization in unadjusted models for wild‐type, BA.1, and BA.5 strains and against the BA.1 strain in adjusted models. Rituximab was associated with lower neutralization for the BA.1 strain in adjusted models, with a similar trend for BA.5. In methotrexate users, there were trends for less neutralization of BA.1 and BA.5 in all unadjusted models, whereas in adjusted models, there was significantly lower neutralization only for the wild type. Three or more doses and Omicron‐specific vaccines were both independently associated with better neutralization ability for all three strains. A COVID‐19 infection within six months before sampling was associated with higher neutralization of wild type and BA.1 in adjusted analyses. Anti–tumor necrosis factor agents were associated with lower neutralization ability for BA.5 in adjusted analyses.

Conclusion

Neutralization responses in immunosuppressed individuals with IMID were durable over time and were augmented by more than three doses and Omicron‐specific vaccines. Less neutralization was seen with certain medications. Our work clarifies the joint effects of vaccine history, infection, and medications on COVID‐19 immunity.

INTRODUCTION

Although we are recovering from the COVID-19 pandemic, SARS-CoV-2 remains a public health threat,¹ particularly for those taking immunosuppressants for immune-mediated inflammatory disorders (IMIDs), such as inflammatory bowel disease (IBD), psoriasis (including psoriatic arthritis [PsO/PsA]), rheumatoid arthritis (RA), spondylarthritis (SpA), and systemic lupus erythematosus (SLE).^2–5 Serological responses (eg, anti-Spike [S] or anti–receptor binding domain [RBD]) to COVID-19 vaccination may be diminished by glucocorticoids, methotrexate, anti–tumor necrosis factor (TNF) agents, rituximab (RTX), and other agents, but little is known regarding how well the antibodies generated can neutralize SARS-CoV-2 and the durability of this response. The Safety and immUnogenicity of Covid-19 vaCcines in systEmic immunE mediated inflammatory Diseases (SUCCEED) project was funded by Canada's COVID-19 Immunity Task Force to help decision-makers regarding vaccination strategies. As people with IMIDs face new waves of COVID infection in 2024 and beyond, it remains extremely important to understand the durability of postvaccination responses (including the ability to neutralize COVID-19 strains), including effects of medications and other variables.

METHODS

In these observational cross-sectional analyses, we analyzed prospectively collected paired samples in adults with IBD, PsO/PsA, RA, AS/SpA, and SLE. Participants were recruited from participating Canadian medical centers from 2021 to 2023. Exclusion criteria included individuals who could not provide consent in English or French or whose last COVID vaccination had been longer than six months before recruitment (unless they planned to be vaccinated shortly). The current analyses focused on the presence of neutralizing antibodies detected in paired sera (pre/post–third or fourth vaccination) collected from participants in Toronto, Hamilton, Québec City, Sherbrooke, and Winnipeg. Participants provided baseline information on demographics, COVID-19 vaccinations (including dates and types), COVID-19 infections, and clinical history (type of IMID, date of diagnosis, medications).

Sera were collected from patients and then batch-processed in the Gingras laboratory in Toronto.⁶ For the neutralization studies, lentivirus particles were generated by co-transfection in HEK293TN cells (LV900A-1, System Biosciences) of the Wuhan Hu-1 sequence with a D614G mutation (wild-type SARS-CoV-2) or the Omicron variants BA.1 and BA.5. Heat-inactivated (30 minutes at 56°C) sera was serially diluted (3-fold for wild-type and 2.5-fold for Omicron variants) over seven dilutions and incubated with the lentiviral particles (for 1 hour at 37°C) before adding the sera to cells (HEK293T-ACE2/TMPRSS2). After 48 hours, luminescence signals were detected with the Bright-Glo Luciferase assay system (catalog E2620, Promega) on an EnVision multimode plate reader (Perkin Elmer). Relative luciferase units were first normalized to positive and negative controls, then a sigmoidal curve was fitted and an ID₅₀ (the ID₅₀ value is the inhibitory dilution at which 50% neutralization is attained) value (50 % neutralization titers) was calculated for each sample.

Our ID₅₀ results were log transformed to create a more linear distribution of the data so that we could use generalized estimating equation (GEE) linear models (to account for the fact that we had two samples per individual). Outcomes were analyzed for the wild-type, Omicron B.A.1, and B.A.5 variants separately. We described the demographics of individuals providing samples and attempted to determine whether production of neutralizing antibodies was negatively associated with time since last vaccine and with baseline use (at study enrollment) of medications, adjusting for sex, age, race/ethnicity, IMID type, and details of past COVID vaccinations and infections.

Multivariate GEE models were created with the log-transformed neutralization outcome as a continuous variable. In our models, the exponentiated beta coefficient for each variable can be interpreted as representing a relative decrease or increase in the outcome; for example, an exponentiated beta coefficient of 0.90 suggests a 10% decrease relative to the reference, and an exponentiated beta coefficient of 1.10 suggests a 10% increase relative to the reference.

This study was approved by the research ethics board of the Research Institute of the McGill University Health Centre and by the research ethics boards of all participating institutions. A data dictionary defining each field in the data set is available upon request from the corresponding author.

RESULTS

A total of 112 participants represented Manitoba (n = 57, 50.9%), Ontario (n = 26, 23.2%), and Québec (n = 29, 25.9%). Just over half (69 of 112) of participants were female with a mean ± SD age of 61.7 ± 14.8 years at the time the second sample was obtained. Most participants (n = 99, 88.4%) were self-described as White/Caucasian; the remainder were indigenous (n = 5, 4.5%) or Black/Asian (n = 4, 3.6%), with four participants preferring not to disclose their race/ethnicity. The types of IMID included RA (n = 54, 48.2%), IBD (n = 27, 24.1%), PsA (n = 19, 17%), SpA (n = 11, 9.8%), and SLE (n = 1, 0.89%). At the time of enrollment into the SUCCEED study, 16 (14.3%) participants were on prednisone and 32 (28.6%) were on a biologic, most often (n = 20) a TNF agent (others included abatacept in 5 cases, RTX in 3 cases, ustekinumab in 2 cases, and 1 case each for vedolizumab and tocilizumab). Ninety-eight of the 112 (86.6%) participants were on a non–biologic disease-modifying antirheumatic drug, most often methotrexate (n = 41) and/or hydroxychloroquine (n = 35), in addition to sulfasalazine (n = 10), leflunomide (n = 5), a JAK inhibitor (n = 4), or azathioprine and 6-mercaptopurine (1 each). The most common vaccines at the first study sample used in our analyses were the monovalent BNT162b2 only (n = 55, 49.1%), followed by combinations of mixed monovalent vaccine (n = 30, 26.8%) and bivalent plus any other combination (n = 23, 20.5%). Of the 23 patients who received a bivalent vaccine, 16 received the Spikevax original/Omicron BA.4–5 vaccine and 7 received the COMIRNATY original/Omicron BA.4–5 vaccine.

Compared with the period of 30 to 120 days postvaccination, subsequent time (either 121–210 or 211+ days after last vaccine dose) was associated with a higher neutralization ability in unadjusted models for all three strains, although in adjusted models, this remained apparent for BA.1 only (adjusted exponentiated coefficient [Expβ] 2.51, 95% confidence interval [95% CI] 1.17–5.37). Figure 1 shows neutralization values related to time since last vaccination. Most individuals (93, 83% of the 112) contributed their samples within 120 days of their last vaccine; thus, the results from the past 120 days rely on data from 19 individuals only (and only 2 of these individuals contributed samples more than 210 days since their last vaccination).

[IMAGE OMITTED. SEE PDF]

Exposure to vaccines formulated against Omicron (in our case, bivalent versions available in 2022–2023) was also associated with better neutralization ability in adjusted models for wild-type and both Omicron strains (Table 1). Three or more doses of any COVID-19 vaccine was also associated with better neutralization ability for all three strains in the unadjusted model. COVID-19 infection within six months of sampling was associated with neutralization of wild-type and both Omicron strains in adjusted analyses.

Table 1 Effects of covariates on neutralization of wild-type and Omicron BA.1 and BA.5 strains: Exponentiated β coefficient (Expβ)^a and 95% confidence intervals (CI).

	Wild-type Expβa (95% CI)	Omicron BA.1 Expβa (95% CI)	Omicron BA.5 Expβa (95% CI)
Crude	Adjustedb	Crude	Adjustedb	Crude	Adjustedb
Female sex	0.91 (0.63, 1.35)	1.20 (0.81, 1.77)	0.71 (0.39, 1.30)	1.09 (0.68, 1.79)	0.80 (0.46, 1.39)	1.15 (0.70, 1.86)
Age, years	1.05 (0.90, 1.23)	1.01 (0.87, 1.17)	1.40 (1.06, 1.88)	1.16 (0.94, 1.42)	1.09 (0.88, 1.36)	0.87 (0.74, 1.03)
White race/ethnicity	1.25 (0.90, 1.73)	1.17 (0.83, 1.68)	2.03 (1.12, 3.71)	1.52 (0.89, 2.61)	1.99 (1.09, 3.67)	1.80 (1.16, 2.83)
Disease (reference: other-psoriasis/psoriatic arthritis, ankylosing spondylitis/spondylarthritis, SLE)
Inflammatory bowel disease	0.90 (0.52, 1.54)	1.48 (0.84, 2.64)	0.37 (0.16, 0.86)	1.20 (0.55, 2.59)	0.44 (0.22, 0.90)	1.46 (0.79, 2.69)
Rheumatoid arthritis	1.11(0.63, 1.93)	1.65 (0.92, 2.94)	0.95 (0.41, 2.20)	1.7 (0.78, 4.10)	0.78 (0.38, 1.60)	1.52 (0.76, 3.03)
Vaccine type (reference: bnt162b2 only)
mRNA1273 only	1.48 (0.99, 2.20)	1.31 (0.77, 2.20)	1.40 (0.76, 2.61)	1.54 (0.58, 4.10)	1.09 (0.53, 2.27)	1.17 (0.55, 2.46)
Mixed monovalent	1.26 (0.83, 1.92)	1.21 (0.81, 1.80)	0.91 (0.52, 1.63)	0.77 (0.49, 1.22)	0.95 (0.53, 1.70)	0.81 (0.49, 1.35)
Omicron-specific	2.41 (1.67, 3.53)	2.34 (1.55, 3.53)	6.55 (3.97, 10.7)	5.00 (2.77, 9.03)	5.58 (3.71, 8.41)	4.90 (2.86, 8.33)
3+ vaccine doses	1.40 (1.02, 1.95)	1.22 (0.85, 1.77)	2.36 (1.49, 3.74)	1.25 (0.76, 2.05)	2.05 (1.32, 3.19)	1.19 (0.77, 1.80)
Days since last COVID vaccine (reference: between 30–120 days)
121–210	1.14 (0.76, 1.72)	0.97 (0.63, 1.48)	2.36 (1.19, 4.66)	1.16 (0.64, 2.12)	2.27 (1.28, 4.01)	1.26 (0.76, 2.08)
211+	2.48 (2.05, 3.00)	1.65 (0.94, 2.86)	2.46 (1.90, 3.22)	2.51 (1.17, 5.37)	1.77 (1.38, 2.29)	2.12 (0.95, 4.71)
Baseline prednisone (yes/no)	1.02 (0.62, 1.70)	0.90 (0.58, 1.40)	1.43 (0.74, 2.77)	1.05 (0.63, 1.73)	1.46 (0.75, 2.86)	1.20 (0.68, 2.12)
Baseline biologic (reference: no biologic)
Anti-tumor necrosis factor agent	0.72 (0.44, 1.20)	0.79 (0.45, 1.38)	0.56 (0.27, 1.17)	0.80 (0.45, 1.43)	0.50 (0.25, 0.99)	0.55 (0.31, 0.99)
Rituximab	0.37 (0.26, 0.54)	0.57 (0.31, 1.05)	0.97 (0.53, 1.75)	0.49 (0.25, 0.95)	0.91 (0.51, 1.65)	0.69 (0.36, 1.31)
Other biologicc	0.93 (0.36, 2.39)	1.09 (0.50, 2.41)	1.95 (0.86, 4.48)	1.73 (0.78, 3.90)	1.04 (0.32, 3.39)	1.12 (0.44, 2.86)
Baseline non-biologic immunomodulators (reference; none)
Hydroxychloroquine	1.12 (0.75, 1.67)	1.17 (0.71, 1.93)	1.48 (0.84, 2.56)	1.25 (0.67, 2.34)	1.30 (0.76, 2.20)	1.32 (0.70, 2.51)
Methotrexate	0.70 (0.48, 1.02)	0.58 (0.39, 0.85)	0.88 (0.51, 1.52)	0.64 (0.38, 1.09)	0.76 (0.45, 1.28)	0.71 (0.43, 1.20)
COVID infection within 6 months before sample	2.32 (1.22, 4.39)	2.83 (1.40, 5.70)	4.01 (0.62, 26.1)	3.97 (1.01, 15.5)	2.66 (0.51, 14.1)	2.89 (1.00, 9.97)

Anti-TNF agents were associated with significantly lower neutralization ability for BA.5 in the adjusted (Expβ 0.55, 95% CI 0.31–0.99) and unadjusted (Expβ 0.50, 95% CI 0.25–0.99) models. Twenty individuals were exposed to TNF inhibitors, including three who had received one Omicron-specific formulation (Spikevax or COMIRNATY original/Omicron BA.4–5). All three individuals developed very high neutralization ability against BA.5 (average 4,336, median 4,585). Among the 17 remaining TNF inhibitor–exposed individuals who did not receive at least 1 bivalent dose, only 2 (11.8% of the 17) had high neutralization against BA.5 (ie, >850 ID₅₀). Among the three individuals exposed to TNF inhibitors who received at least one bivalent dose, average ID₅₀ neutralization results were 4,336 (median 4,585), whereas for those who had not received any vaccination against Omicron variants, the mean ID₅₀ level was 395 (median ID₅₀ level was 128).

Rituximab was associated with lower neutralization ability for the wild-type strain in the unadjusted model and BA.1 strains in the adjusted model (Expβ 0.49, 95% CI 0.25–0.95) with a similar trend in the adjusted model for BA.5. For methotrexate, there were trends for less neutralization ability for all three strains (albeit with quite wide 95% CIs in the adjusted models) and a significantly lower neutralization for the wild type in the adjusted model (Expβ 0.58, 95% CI 0.39–0.85). There were not enough individuals on JAK inhibitors to comment on these drugs. Prednisone use at cohort entry was not clearly associated with neutralization ability.

Additional findings are shown in Table 1. Of note, persons of White race tended to have higher neutralization ability, particularly for Omicron variants (eg, Expβ for BA.5 was 1.80, 95% CI 1.16–2.83). Our variable for age did not show consistent associations with neutralization, although the crude Expβ estimate for BA.1 neutralization was elevated (Expβ 1.40, 95% CI 1.06–1.88). Conversely, crude Expβ were lower in individuals with IBD for both BA.1 (Expβ 0.37, 95% CI 0.16–0.86) and BA.5 (Expβ 0.44, 95% CI 0.22–0.90).

DISCUSSION

Data demonstrating that immunosuppressed individuals may mount insufficient serological responses to immunization after two doses of a messenger RNA (mRNA) vaccine⁶ has given rise to recommendations for three doses as the primary series of COVID-19 vaccination in immunosuppressed individuals. Our analyses extended these findings to the outcome of neutralization ability, not only to the wild-type variant but also to Omicron strains. Interestingly, mean neutralization ability in samples taken more than 120 days after the last vaccine was higher for wild-type and BA.1 strains, with a similar trend (just under the significance level in terms of 95% CI) for BA.5. This is in keeping with what others have suggested regarding the generation of durable neutralization responses. At the same time, infection within the previous six months was in multivariant analyses associated with neutralization ability for the wild-type and Omicron strains. In our analyses, and as expected, bivalent vaccines were associated with more neutralizing antibody response to both wild-type and Omicron strains, compared with vaccines that were developed based on the wild-type strain. Ours is the first Canadian multicenter assessment of medication-exposed patients with IMIDs to specifically show these results.

Regarding medication exposures, it has been previously reported that serological responses to anti-S and anti-RBD after vaccination may be diminished in individuals on methotrexate, anti-TNF agents, RTX, and other immunosuppressants.^7–10 For some of these drugs, including anti-TNF agents, humoral immunity against COVID-19 vaccination has been shown not only to be lower but to decline earlier, as well.¹¹ Interestingly, one study of IMID patients, though demonstrating lower anti-S titers for these drug exposures, did not detect clear differences in neutralization ability.¹² However, that study examined only neutralization against the wild-type strain (not Omicron strains). Similarly, Mahil et al did not detect adverse effects for these medications on neutralization ability but studied only pre-Omicron variants.¹³ In our study, which evaluated neutralization ability against wild-type and Omicron strains, the effects of anti-TNF agents and RTX were easier to demonstrate for some variants than others. Some anti-TNF exposures have also been shown to be associated with breakthrough infections.¹⁴ Our study was not designed to look at this outcome.

In our study, RTX was associated with lower neutralization ability for the wild-type strain in the unadjusted model and BA.1 strains in adjusted models with a similar trend in the adjusted model for BA.5. Although previous studies raised concerns about RTX's depression of immune response to COVID vaccination, most had not focused solely on IMID, nor assessed responses beyond the first year of vaccination.¹⁵ We did not have enough data to examine the timing of RTX infusions with respect to vaccination or subsequent neutralization ability. RTX is primarily used in RA and sometimes SLE (but not in IBD or PsA/PsO or SpA), whereas anti-TNF agents are often used in both RA and IBD (as well as PsA/PsO or SpA), but not in SLE. Thus, the benefit of combining data across a number of different IMIDs was a strength of our study. Prednisone use at cohort entry was not clearly associated with neutralization ability, but there were relatively few individuals on this agent in our data.

We found that persons of White race tended to have higher neutralization ability, particularly for Omicron variants. Some studies in non-IMID populations have suggested lower neutralization ability in White participants after SARS-CoV2 vaccination, including one study of United States health care workers.¹⁶ However, we had a limited number of patients who were not White, including only five indigenous participants, four Black/Asian participants, and four participants who preferred not to disclose race/ethnicity. Thus, our results with respect to race/ethnicity should be interpreted with caution.

Our variable for age did not show consistent associations with neutralization, although the crude Expβ estimate for BA.1 neutralization was elevated, possibly reflecting earlier uptake of vaccination (and number of vaccinations) in seniors. Conversely, crude Expβ were lower in individuals with IBD for both BA.1 and BA.5, potentially because of anti-TNF exposures or other factors (IBD samples tend to come from the calendar year period and thus have fewer vaccinations and past COVID infections). Our study has many strengths. We assessed individuals from a spectrum of IMID types using different therapies across multiple Canadian centers (“real-world study”). We evaluated the joint effects of demographics, drug, vaccine history, and COVID-19 infection, which only some previous studies were able to do. Moreover, we assessed neutralization to wild-type and Omicron variants, in contrast to many previous studies that assessed only neutralization to the pre-Omicron variant. Moreover, we provide estimates of responses to newer formulations (specifically targeting Omicron variants), which provides novel information. However, there are some potential limitations to acknowledge. First, because we did not focus on single IMIDs, individuals were on a diverse number of therapies that led to imprecision for some drug effect estimates. We simply could not address the impact of advanced therapies other than anti-TNF and RTX because of small sample sizes. Furthermore, because most of the samples were collected within 120 days of the last vaccine, we had limited ability to determine effects of time since last vaccination beyond this point.

In conclusion, this multicenter study of neutralization ability against wild-type and Omicron strains in IMIDs identified opportunities to personalize decision-making for immunosuppressed populations. Neutralization responses in immunosuppressed individuals with IMIDs were durable over time and were augmented by three or more doses and Omicron-specific vaccines. We confirmed earlier findings suggesting more neutralization ability with three or more vaccines in immunosuppressed IMID individuals and less neutralization with certain medications. Our study further adds valuable information on the role that previous infections play in hybrid immunity and the beneficial effects on neutralization ability that newer vaccine formulations (targeting Omicron variants) have produced. Additional studies are underway in a larger IMID sample regarding the kinetics of the development of neutralization ability (both in terms of time to develop a response and waning) and further evaluation of vaccine and infection history, as well as drug exposures.

AUTHOR CONTRIBUTIONS

All authors were involved in drafting the article or revising it critically for important intellectual content, and all authors approved the final version to be published. Dr Bernatsky had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Study conception and design

Dayam, Hitchon, Chandran, Fortin, Boire, Bowdish, Gingras, Flamand, Larché, Colmegna, Bernatsky.

Acquisition of data

Hitchon, Chandran, Fortin, Boire, Gingras, Larché, Lee, Pereira, Bernstein, Turnbull, Bernatsky.

Analysis and interpretation of data

Habib, Dayam, Hitchon, Chandran, Fortin, Boire, Bowdish, Gingras, Flamand, Larché, Colmegna, Lukusa, Lalonde, Bernatsky.

References

COVID‐19 remains a persistent public health threat. World Health Organization. Accessed April 5, 2024. https://www.emro.who.int/media/news/covid-19-remains-a-persistent-public-health-threat.html

Eder L, Croxford R, Drucker AM, et al. COVID‐19 hospitalizations, intensive care unit stays, ventilation, and death among patients with immune‐mediated inflammatory diseases compared to controls. J Rheumatol 2022;49(5):523–530.

Widdifield J, Eder L, Chen S, et al. COVID‐19 vaccination uptake among individuals with immune‐mediated inflammatory diseases in Ontario, Canada, between December 2020 and October 2021: a population‐based analysis. J Rheumatol 2022;49(5):531–536.

Eder L, Croxford R, Drucker AM, et al. Understanding COVID‐19 risk in patients with immune‐mediated inflammatory diseases: a population‐based analysis of SARS‐CoV‐2 testing. Arthritis Care Res (Hoboken) 2023;75(2):317–325.

Widdifield J, Kwong JC, Chen S, et al. Vaccine effectiveness against SARS‐CoV‐2 infection and severe outcomes among individuals with immune‐mediated inflammatory diseases tested between March 1 and Nov 22, 2021, in Ontario, Canada: a population‐based analysis. Lancet Rheumatol 2022;4(6):e430–e440.

Dayam RM, Law JC, Goetgebuer RL, et al. Accelerated waning of immunity to SARS‐CoV‐2 mRNA vaccines in patients with immune‐mediated inflammatory diseases. JCI Insight 2022;7(11): [eLocator: e159721].

Sakuraba A, Luna A, Micic D. Serologic response to coronavirus disease 2019 (COVID‐19) vaccination in patients with immune‐mediated inflammatory diseases: a systematic review and meta‐analysis. Gastroenterology 2022;162(1):88–108.

Deepak P, Kim W, Paley MA, et al. Effect of immunosuppression on the immunogenicity of mRNA vaccines to SARS‐CoV‐2: a prospective cohort study. Ann Intern Med 2021;174(11):1572–1585.

Alexander JL, Kennedy NA, Ibraheim H, et al; VIP study investigators. COVID‐19 vaccine‐induced antibody responses in immunosuppressed patients with inflammatory bowel disease (VIP): a multicentre, prospective, case‐control study. Lancet Gastroenterol Hepatol 2022;7(4):342–352.

Edelman‐Klapper H, Zittan E, Bar‐Gil Shitrit A, et al; REsponses to COVid‐19 vaccinE IsRaeli IBD group (RECOVER). Lower serologic response to COVID‐19 mRNA vaccine in patients with inflammatory bowel diseases treated with anti‐TNFα. Gastroenterology 2022;162(2):454–467.

Christensen IE, Jyssum I, Tveter AT, et al. The persistence of anti‐Spike antibodies following two SARS‐CoV‐2 vaccine doses in patients on immunosuppressive therapy compared to healthy controls‐a prospective cohort study. BMC Med 2022;20(1):378.

Wieske L, van Dam KPJ, Steenhuis M, et al; T2B! Immunity against SARS‐CoV‐2 study group. Humoral responses after second and third SARS‐CoV‐2 vaccination in patients with immune‐mediated inflammatory disorders on immunosuppressants: a cohort study. Lancet Rheumatol 2022;4(5):e338–e350.

Mahil SK, Bechman K, Raharja A, et al. Humoral and cellular immunogenicity to a second dose of COVID‐19 vaccine BNT162b2 in people receiving methotrexate or targeted immunosuppression: a longitudinal cohort study. Lancet Rheumatol 2022;4(1):e42–e52.

Lin S, Kennedy NA, Saifuddin A, et al; CLARITY IBD study. Antibody decay, T cell immunity and breakthrough infections following two SARS‐CoV‐2 vaccine doses in inflammatory bowel disease patients treated with infliximab and vedolizumab. Nat Commun 2022;13(1):1379.

Boekel L, Steenhuis M, Hooijberg F, et al. Antibody development after COVID‐19 vaccination in patients with autoimmune diseases in the Netherlands: a substudy of data from two prospective cohort studies. Lancet Rheumatol 2021;3(11):e778–e788.

Ahluwalia P, Vashisht A, Singh H, et al. Ethno‐demographic disparities in humoral responses to the COVID‐19 vaccine among healthcare workers. J Med Virol 2023;95(9): [eLocator: e29067].