Serological Responses to the First Three Doses of SARS-CoV-2 Vaccination in Inflammatory Bowel Disease: A Prospective Cohort Study

Background

Individuals with inflammatory bowel disease (IBD) may have an increased risk of infection and complications of coronavirus disease 2019 (COVID-19) because of disease activity and/or an immunocompromised state due to their medical therapy. Highly effective vaccines against Severe Acute Respiratory Syndrome Coronavirus 2 (SARS-CoV-2) have been made widely available to reduce disease severity of COVID-19.¹ The most commonly used vaccines in Canada have been: Pfizer-BioNTech BNT162b2 mRNA (Comirnaty®), NIH-Moderna mRNA-1273 (Spikevax®), and AstraZeneca ChAdOx1 nCoV-19 (Vaxzevria®) adenovirus vector vaccine, which were initially administered as two-dose regimens.^2-4

Several studies in the IBD population have demonstrated adequate serological response to the two-dose regimen of vaccines.⁵ However, a reduced antibody response has been reported for those on certain treatments, including: tumor necrosis factor antagonists (anti-TNF), immunomodulators, tofacitinib, oral corticosteroids, and combinations of immunosuppressive therapies.^6-11 Moreover, the decay of antibody levels has been shown in the IBD population, with therapies such as anti-TNF associated with a more rapid decline.⁷ Consequently, gastroenterology organizations and public health officials have recommended three-dose vaccine regimens for individuals with IBD.¹² Despite recommendations, uptake of 3^rd dose vaccines remains low in the IBD population.¹³

We examined the serological response following the 1^st, 2^nd, and 3^rd dose of SARS-CoV-2 vaccines in those with IBD; the decay of antibodies over time; and the factors, including medications, associated with antibody titres in those with IBD.

Methods

Results

Table 1 describes the distribution of demographic, disease, and vaccine-related characteristics for the entire cohort (n = 496) and for individuals with serological data 1–8 weeks following 1^st dose (n = 250), 1–8 weeks following 2^nd dose (n = 315), 8+ weeks following 2^nd dose (n = 237) and 1+ weeks following 3^rd dose vaccination (n = 232).

Seroconversion rates were 81.6% (1^st dose, 1–8 weeks), 98.7% (2^nd dose, 1–8 weeks), 94.9% (2^nd dose, 8+ weeks), and 99.6% (3^rd dose, 1+ weeks) with 100% seroconversion for those with a prior COVID-19 infection for all dose categories (Table 1).

GMT levels significantly increased (p<0.001) from 1–8 weeks after 1^st dose (1814 AU/mL [95% CI: 981, 2648 AU/mL]) to 1–8 weeks after 2^nd dose (9229 AU/mL [95% CI: 7871, 10588 AU/mL]), but fell significantly (p<0.001) to 3002 AU/mL (95% CI: 2222, 3783 AU/mL) 8+ weeks after 2^nd dose vaccination (Table 1, Figure 1). Following 3^rd dose vaccination, antibody levels increased significantly compared to all vaccine categories (p<0.001) to 14639 AU/mL (95% CI: 12846, 16472 AU/mL). The distribution of individual antibody concentrations per vaccine dose category is displayed in Figure 1.

• Download figure
• Open in new tab

Figure 1.

Anti-SARS-CoV-2 spike antibody concentration per vaccine category. Black points represent GMTs while narrow black bars represent bounds of 95% CI associated with each GMT. Solid blue line represents threshold for positive seroconversion (50 AU/mL). Serological data from this study are available open access on an online interactive ShinyApp: https://kaplan-gi.shinyapps.io/COVID_Serology/

Age at time of serology was independently associated with decreased log-transformed anti-S concentration for all vaccine dose groups except post-3^rd dose. FC values indicate factor of log anti-S concentration per decade increase in age, for example log anti-S concentration decreased by 0.71 (95% CI: 0.64, 0.80) per decade increase in age for samples taken 8+ weeks post-2^nd dose vaccine. Three-dimensional plots displaying anti-S concentration by age and days after vaccination indicate decreasing titres with increasing age and time after vaccination for 2^nd dose, whereas for 3^rd dose, titres only decrease with time after vaccination (Video 1 and Video 2). Age >65 years was associated with lower GMT for all vaccine dose groups except post-3^rd dose (Supplemental Table 1).

For each vaccine dose category, patients on corticosteroids had the lowest GMT compared to other medication classes (Table 2, Supplemental Table 2). Multivariable linear regression analyses demonstrated a significant association between corticosteroid use and decreased log-transformed anti-S concentration compared to individuals not on immunosuppressive medication, across all vaccine dose categories (1–8 weeks post-2^nd dose FC: 0.07 (95% CI: 0.02, 0.20). Anti-TNF monotherapy and combination therapy were also predictors of decreased anti-S titres for all vaccine dose categories except post-3^rd dose vaccination (Table 2). The FC in log anti-S concentration compared to the no immunosuppressive medication group for each medication is displayed in Table 2. For example, at 1–8 weeks post-2^nd dose, anti-TNF monotherapy was associated with a decrease in log anti-S concentration of 0.56 (95% CI: 0.33, 0.94) compared to individuals without immunosuppressive medications.

Prior COVID-19 history was significantly associated with increased log anti-S concentration across all vaccine dose categories with the largest FC associated with post 1^st dose responses (FC=21.91 [95% CI: 16.37, 54.64]) and the lowest impact associated with post 3^rd dose responses (FC=1.97 [95% CI: 1.22, 3.20]). Delayed 2^nd dose vaccination schedule (>6 weeks between 1^st and 2^nd doses) was associated with increased log anti-S concentration 8+ weeks following 2^nd dose vaccination (p<0.001). However, the effect of delayed 3^rd dose vaccination schedule (>18 weeks between 2^nd and 3^rd doses) was not observed.

Antibody titres decreased significantly (p<0.001) following 2^nd dose vaccination; Figure 2a demonstrates an inverse trend between anti-S antibody concentrations and days after 2^nd dose vaccination. Figure 2b. presents a similar trend for post-3^rd dose results, with overall lower anti-S antibody concentrations for serology samples taken further from the 3^rd dose vaccination date.

• Download figure
• Open in new tab

Figure 2a.

Anti-S concentration per days after dose vaccination. Figure 2b. Anti-S concentration per days after 3^rd dose vaccination. Blue points represent individuals with prior SARS-CoV-2 infection. Solid red lines represent line of best fit for linear relationship between anti-S concentration (AU/mL) and time after vaccination. Solid blue lines represents threshold for positive seroconversion (50 AU/mL). Serological data from this study are available open access on an online interactive ShinyApp: https://kaplan-gi.shinyapps.io/COVID_Serology/

Multivariable regression analyses identified a significant relationship between weeks after vaccination and log anti-S concentration for 1–8 weeks post-2^nd dose (FC: 0.82 [95% CI: 0.75, 0.89] per week), 8+ weeks post-2^nd dose (FC: 0.92 [95% CI: 0.90, 0.94] per week), and 1+ weeks post-3^rd dose (FC: 0.88 [95% CI: 0.84, 0.92] per week) groups. Antibodies decayed by a factor of 0.88 or 12% per week (95% CI: 8% to 16%) following the 3^rd dose of vaccine.

The mean difference in antibody concentration from post 1–8 weeks to 8+ weeks after 2^nd dose (n=124) was -7668 AU/mL (95% CI: -10055, -5281 AU/mL, p<0.001). After stratifying by prior SARS-CoV-2 infection, differences were only significant for individuals without prior infection (−6970 AU/mL [95% CI: -8999, -4941 AU/mL]). The individual rate of decay in individuals without prior SARS-CoV-2 infection was determined to be 396 AU/mL per week (95% CI: 277, 516 AU/mL per week).

GMT following post-3^rd dose vaccination increased with a mean difference in antibody titres between post-3^rd dose and 1–8 weeks post-2^nd dose vaccination (n = 153) of 8148 AU/mL (95% CI: 6155, 10141 AU/mL, p<0.001). Within-individual comparison following 8+ weeks post-2^nd dose and post-3^rd dose vaccination (n = 99) also demonstrated that GMT following post-3^rd dose vaccination increased significantly with a mean difference of 12400 AU/mL (95% CI: 9281, 15518 AU/mL) (Table 3).

Within-individual comparison of multiple 3^rd dose vaccination serology values (n = 28) demonstrated a mean decrease in individual anti-S concentration of 8057 AU/mL (95% CI: 2303, 13811 AU/mL) between earlier and later collection dates. The median duration between serology collection dates was 13.5 weeks (IQR: 11.1–15.7 weeks). The individual rate of decay following 3^rd dose vaccination in individuals without prior infection to SARS-CoV-2 was determined to be 483 AU/mL per week (95% CI: 237, 729 AU/mL per week).

Sensitivity analyses comparing paired data between vaccine dose categories did not indicate any differences in statistical significance. Sensitivity analyses modelling age as a binary variable (18–65 years versus >65 years) also demonstrated significant associations with increased age (>65 years) and decreased antibody response for post-1^st dose, 1–8 weeks post-2^nd dose, and 8+ weeks post-2^nd dose. Likewise, there was no association for post-3^rd dose responses (p=0.579) (Supplemental Table 1). Subgroup analyses excluding non-mRNA vaccination and vaccine mixing did not impact the significance of covariate associations with log anti-S concentration. Subgroup models for post 1^st dose and 1–8 weeks post-2^nd dose responses indicate significantly decreased log anti-S concentration for Pfizer compared to Moderna (FC: 0.41 [95% CI: 0.24, 0.70]; FC: 0.51 [95% CI: 0.34, 0.78]) (Supplemental Table 3). Supplemental Table 4 reports antibody levels in the World Health Organization BAU/ml.

Discussion

A three-dose SARS-CoV-2 vaccination schedule for individuals with IBD has been proposed based on impaired serological responses to two-dose regimens.^12,21,22 Our study demonstrates near complete seroconversion to 3^rd dose vaccination with significantly higher antibody titres when compared to 2^nd dose vaccination. Third dose antibody responses are consistently high across all IBD therapies for maintenance of remission including biologic and immunomodulator therapies. In contrast, those with IBD taking prednisone at the time of receiving a vaccine mounted lower antibody responses following all three doses. Despite robust initial antibody responses to vaccination, our study indicates decay of antibody titres over time following vaccination for both 2^nd and 3^rd doses. While these data support widespread implementation of a three-dose vaccine regimen, future studies are necessary to determine if additional doses are required to maintain sufficient antibody levels over time.

Our findings indicate a limited antibody response to a single dose of a SARS-CoV-2 vaccine, but consecutively stronger responses following each additional dose. Only 81.6% of individuals with IBD seroconverted following the 1^st dose of a SARS-CoV-2 vaccine. However, 98.7% of participants seroconverted following two doses with a corresponding five-fold increase in mean antibody titres. Similarly, a meta-analysis in the IBD population showed an increase from 49.2% to 93.7% seroconversion after 1^st and 2^nd doses, respectively.⁵ Our study demonstrated a substantial increase in serological response to a 3^rd dose of a vaccine, which was similar to an American cohort of IBD patients that demonstrated a robust antibody response following an extra dose from the initial vaccine series.²³ Consistent seroconversion and antibody titres responses across different geographic jurisdictions and using different antibody assays highlights the replicative external validity of the strong antibody response to three dose regimens of the SARS-CoV-2 vaccines in the IBD population.

While considerably large serological responses were observed immediately after 2^nd and 3^rd doses, the linear regression models showed that antibody titres diminished over time after vaccination. Within-individual analyses also indicated that antibody levels decayed at an approximate rate of 400 AU/mL per week after 2^nd dose and 500 AU/mL per week after 3^rd dose. Decay of antibody titres following vaccine doses has been previously reported;^7,24 however, 3^rd dose antibody titres diminish from a considerably higher baseline as compared to the 2^nd dose vaccine. The threshold level for antibody titres to offer protection from an infection to SARS-CoV-2 is not known. A prior study in the UK demonstrated that following the 2^nd dose of the vaccine that a 10-fold increase in antibody titres was associated with 0.8-fold decrease in breakthrough infections.⁷ Future studies are necessary to establish the antibody level necessary to confer protection to SARS-CoV-2 following three vaccine doses.

Older age was associated with decreased antibody titres after 1^st and 2^nd doses in prior studies,^6-8 which is clinically relevant as seniors with IBD are at highest susceptibility to severe COVID-19.²⁵ A prior study showed that antibody titres fell by 21% per decade of life following the 2^nd dose of vaccine.⁶ Similarly, our models also indicated decreases in antibody concentration per decade of life across 1^st and 2^nd doses. For example, immediately after 2^nd dose of the vaccine, antibody titres were 12% lower per decade of advancing age (Video 1). However, our study presents novel data on the effect of age on 3^rd dose responses and indicates no change in antibody titres per decade of advancing age (FC: 0.91 [95% CI: 0.82, 1.01]). These data highlight the fact that seniors with IBD mount lower antibody responses after a two-dose regimen but mount comparably high antibody responses following a three-dose regime; thus, the importance of a 3^rd dose for sufficient immune protection in seniors with IBD.

We assessed infection to SARS-CoV-2 prior to each dose of vaccine by verifying either a PCR-confirmed molecular diagnosis and/or positive serosurveillance with nucleocapsid antibodies. Our data were consistent with prior studies showing that COVID-19 infection before vaccination increases antibody levels following each dose of the vaccine.^7,8 Whereas prior infection induced a 33-fold increase in 1^st dose antibody responses, exposure to SARS-CoV-2 diminished to only a 2.0-fold increase for 3^rd dose responses. We used a lower threshold for a positive anti-N level (i.e., 0.7) than the manufacturer recommended level of 1.4, which is supported by serosurveillance studies of SARS-CoV-2.²⁶

Our data are consistent with previous studies in demonstrating diminished antibody responses following two doses of a SARS-CoV-2 vaccine for patients on anti-TNF therapy, combination immunosuppressive therapies, and corticosteroids.^{6-11,24,27,28} In contrast, our study showed that the 3^rd dose of the vaccine yielded high antibody titres across all classes of medications used for maintenance of remission in IBD. However, administration of vaccine while concurrently taking prednisone yielded lower antibody response following 1^st, 2^nd, and 3^rd dose of the vaccine. While prednisone was associated with diminished responses, mean antibody titres following 3^rd dose vaccination were still 2.5 times higher than those more than eight weeks following 2^nd dose in patients on oral corticosteroids. Findings across medication class should be interpreted cautiously where small sample size reduces the precision of estimates. For example, small sample size limited our ability to analyze the effect of prednisone across different dosages (e.g., above and below 20 mg). While our data highlight that all individuals with IBD should receive a 3^rd dose of the vaccine, regardless of medication class, those on prednisone at the time of receiving a vaccine may ultimately require four doses.

Several limitations of our study should be considered. Due to a low number of patients with breakthrough COVID-19 infections, we were unable to assess the effectiveness of SARS-CoV-2 vaccination or evaluate whether antibody titres were associated with breakthrough infections. In addition, we did not assess other immune responses to SARS-CoV-2 vaccination, including neutralizing antibodies or T-cell immunity, which may serve as more functional proxies of immune protection over antibody titres.^24,29 Due to logistical constraints in data collection during the pandemic, we were unable to obtain samples corresponding to all vaccine doses for all patients, resulting in semi-independent vaccine dose groups. While a community-based recruitment strategy involving all gastroenterologists in the Calgary zone was utilized, population-based sampling was not possible. Further, we did not compare those with IBD to healthy controls; however, a recent serology study performed in the general population within the same jurisdiction and using the same antibody assay demonstrated higher antibodies than our IBD population after 1^st and 2^nd dose of the vaccine.³⁰ We used the Abbott Architect SARS-CoV-2 IgG II Quant assay, but also converted antibody levels to the World Health Organization BAU/ml (Supplementary Table 4) to support standardization across assays.^17,18

Timing and type of vaccination was variable due to changing vaccine availability in Canada. We conducted a sensitivity analysis comparing scheduled versus delayed vaccine timing. Delayed vaccine timing was associated with increased antibody responses following 2^nd dose vaccination; however, timing of vaccine schedule did not influence antibody levels after the 3^rd dose. Also, most participants received mRNA vaccines, with only 4.0% and 2.2% of patients receiving ChAdOx1 nCoV-19 for 1^st and 2^nd dose vaccination, respectively. Several participants also received mixed doses of mRNA and ChAdOx1 nCoV-19 or mixed doses of differing mRNA vaccines. However, our sensitivity analysis restricting the sample to those with the same mRNA vaccine across all doses confirmed the consistency of the interpretation of the data.

Our study demonstrated novel findings surrounding antibody responses to 3^rd dose vaccination, which indicate the importance toward maintaining humoral immunity across all medication classes and age groups. The one exception was individuals with IBD who were vaccinated while on prednisone, who may require a 4^th dose of vaccine in the future. Considering the significant decay of antibodies following a two-dose regimen and robust response following the 3^rd dose, healthcare providers and public health officials should prioritize communication of the necessity of a three-dose vaccine regimen to ensure sufficient serological protection to patients with IBD. Finally, future studies should evaluate the decay of antibodies following the 3^rd dose to establish whether a fourth dose is ultimately necessary for most individuals with IBD.