Antibody response to the COVID-19 ChAdOx1nCov-19 and BNT162b vaccines after temporary suspension of DMARD therapy in immune-mediated inflammatory disease: an extension study (RESCUE 2)

Introduction

The Omicron variants during the SARS-CoV-2 pandemic has continued to pose a risk to patients with immune-mediated inflammatory disease (IMID) despite vaccination. Worldwide, the vaccine roll out continues and boosters have been administered to improve the durability of humoral immunity. There is growing evidence that patients with IMID on disease-modifying antirheumatic therapy (DMARD) mount a delayed and reduced immune response to COVID-19 vaccination which may result in breakthrough infections.1–3 Recent strategy studies show that the reduced vaccine antibody response can be mitigated by withholding DMARD therapy contemporaneous to COVID-19 vaccination or by the administration of additional vaccine booster doses.4–8 Antibodies elicited by natural COVID-19 infection and via vaccination have been observed to decay rapidly in the first few months after infection in healthy individuals, however, the durability of humoral immunity after vaccination in patients with IMID on DMARD therapy remains less well studied.9–11 We recently reported the results of a multicentre three-arm randomised clinical trial comparing the immunogenicity of the ChAdOx1nCov-19 (AZ) and BNT162b (Pfizer) vaccines in patients with IMID compared with controls. Here, we describe immunogenicity data six months post first vaccination and 4 weeks post booster (third) vaccination with the Pfizer or mRNA-1273 (Moderna) vaccine on participants in the same trial and in addition we aim to determine the duration of antibody persistence for each of the DMARD groups.

Methods

The eligibility criteria and SARS-CoV-2 IgG assay evaluation have been previously reported.6 Serum SARS-CoV-2 IgG detection and titres against the S1/2 proteins were measured at baseline, 3–4 weeks post first vaccination and 4 weeks post second vaccination. Participants with IMID were randomised to continue or withhold their DMARD in each of the vaccine groups following the first dose only and the same intervention was then applied following the third (booster) dose.6 To summarise, in participants on weekly methotrexate, the vaccination was timed on the day the dose was due however the methotrexate dose was paused on the day of vaccination for two cycles. Participants on daily DMARDs withheld therapy for 1 week starting on the day of first vaccination. For those on bDMARDs, therapy was delayed by 1 week following their usual injection cycle (eg, for bDMARD administered every 2 weeks, the vaccination was timed at the end of the 2 weeks and then restarted 1 week later leaving an interval of 3 weeks). Participants who contracted COVID-19 infection prior to the third vaccine were excluded in the analysis. The participants were stratified according to the vaccine received and intervention group allocated during the first two vaccinations (table 1). The drugs used in each of the DMARD classes are listed in online supplemental table 1.

Statistical analyses and graphs were performed using either STATA 17.0 (StataCorp) or RStudio. Fisher’s exact test was used to compare seroconversion rates between the various DMARD and control groups while the Wilcoxon-Mann-Whitney U test was used to compare antibody titres between the groups. We conducted univariate logistic regression to assess predictive factors for developing protective SARS-CoV2 IgG antibody titre levels. Multivariate logistic regression was then performed to generate adjusted estimates for factors, which were found to be significant on univariate analysis.

We conducted simple linear regression using log-transformed IgG titre levels as the outcome variable and time since second vaccine dose administration as the explanatory variable to investigate the rate of antibody decay among the cohort studied.

Results

Decay rate of SARS-CoV-2 IgG after vaccination in IMID and controls

The durability of humoral immunity after the second vaccine dose and just prior to the booster dose showed that the SARS-CoV-2 Ab titres were lower with the AZ compared with Pfizer vaccine in both the IMID and control groups. Using a SARS-CoV-2 Ab titre cut-off of 7.0 U/mL as determined previously to equate to a neutralising antibody titre of 50 which was found to be the minimum correlate of protection against SARS-CoV-2 in rhesus macaques, the mean time interval until the loss of protection in the IMID group was 61 days [log10(7)−1.33/−0.008] for the AZ, and 137.5 days [log10(7)−1.67/−0.006] for the Pfizer vaccine6 12 (figure 1a,b).12 Within the intervention groups the durability of humoral response was best in the withhold group for the AZ vaccine (p=0.001) while it was better in the control group for the Pfizer vaccine (p=0.013). In the control group, the number of days before the antibody levels reached below 7.0 U/mL was longer at 62.1 days [log10(7)−1.51/−0.0107] and 209.7 days [log10(7)−1.60/−0.0036] for the AZ and Pfizer vaccines, respectively. The number of patients within each DMARD class who had non protective antibody levels at 6 months post second dose was significantly higher in the tsDMARD group (p=0.026) (figure 2).

Figure 1

Request permissions

Figure 2

Request permissions

When stratified across the different DMARD groups, the AZ vaccine time interval following the second vaccination when the SARS-CoV-2 Ab titres dropped below 7.0 U/mL were 68.3, 71.8 and 64.0 days in the csDMARD, bDMARD and tsDMARD groups, respectively. For the Pfizer vaccine, it was 185.5, 137.5 and 116.0 days csDMARD, bDMARD and tsDMARD groups, respectively (table 4, online supplemental figure 1).

Table 4

View inline•Open as popup

SARS-CoV-2 IgG titre decay rates in IMID versus controls

Factors associated with the magnitude of the humoral response to the AZ and mRNA vaccines

Multivariate modelling of SARS-CoV-2 IgG Ab titres at 6 months postvaccination show increasing age was associated with lower odds of having an IgG level >7.0 U/mL (OR 0.95 (95% CI 0.93 to 0.97), p=0.000). In addition, relative to csDMARDs, patients on tsDMARDs were at a lower odds of having an IgG level >7.0 U/mL at 6 months (OR of 0.38 (95% CI 0.17 to 0.88, p=0.023)). The Ab titres were higher in the group that withheld DMARD therapy (OR 2.89 (95% CI 1.42 to 5.87, p=0.003)) and in those who had the mRNA vaccines (OR 3.43 (95% CI 1.84 to 6.41), p=0.000). After adjusting for confounding, only the peak IgG antibody titre post second vaccination was significant in having an IgG level >7.0 U/mL at 6 months (OR 1.07 (95% CI 1.04 to 1.10), p=0.000)) (online supplemental table 3).

Discussion

Our study shows that there is a significant drop in SARS-CoV-2 IgG Ab titres between the second vaccine dose and the 6 month booster dose in all study arms, however, the seroconversion rates remained similar to controls. Patients with IMID on tsDMARDS who continued with therapy, however, had a lower SARS-CoV-2 IgG Ab titre after 6 months compared with those who withheld therapy and the control group (p=0.010), which is in keeping with another study showing that JAK inhibitors result in a significantly reduced humoral response to vaccination.13 Following the booster vaccination with an mRNA vaccine, the seroconversion rates were 100% in all groups with a corresponding rebound in SARS-CoV-2 IgG Ab levels that even exceeded levels following the second vaccine dose. This was especially in those who received the AZ vaccination initially suggesting that the biological ceiling had not yet been reached.6 Both seroconversion rates and SARS-CoV-2 IgG Ab levels were higher in the Pfizer group for the 6 months and post booster doses. Subsequent vaccine dosing with an mRNA vaccine mitigated the effect of withholding or continuing therapy making it a less important factor in achieving an adequate humoral response which is consistent with previous studies.14

The most important factor in influencing the SARS-CoV2 level at 6 months was the peak antibody level following the second vaccination of the primary vaccine regimen. Studies examining the antibody decay kinetics in participants following COVID-19 infection showed that the height of the peak is dependent on disease severity with antibodies for IgG remaining above the positivity threshold for up to 500 days.15–17 Following vaccination our study showed that SARS-CoV2-Ig Ab persistence (≥7.0 U/mL) was 62 days for the AZ and 210 days for the Pfizer vaccine. Although the antibody levels may persist for longer below 7.0 U/mL, this relative discrepancy suggests that vaccination provides a shorter duration of humoral coverage than natural infection which is consistent previous studies showing a higher risk for breakthrough infection in the vaccinated compared with natural infection.18 19 For patients with IMID on DMARDs the SARS-CoV-Ig Ab persistence was even shorter which supports the need for regular boosting in this cohort especially for those on tsDMARDs.

Longitudinal vaccine responses following the Pfizer and Moderna vaccines have shown that the reduction in SARS-CoV2 IgG Ab to be relatively log-linear.9 11 Although our study measured the SARS-CoV2 Ab levels at two time points only, we found the antibody decay rates to be highest in participants on tsDMARD with a larger gradient in both the AZ and Pfizer vaccines groups compared with patients on csDMARD and bDMARDs. The decay rates were also mostly higher in the Pfizer group on DMARD therapy compared with the AZ group on DMARD therapy. However, due to the higher initial SARS-CoV-2 IgG antibody titres in the Pfizer group, the time for the Ab level to fall below the cut-off was still longer in the Pfizer group. The SARS-CoV-2 IgG antibody reduction rate being dependent on the initial antibody level is supported by another study with higher initial values yielding faster decay rates.20 This observation, however, was reversed in the control group whereby the decay rate was higher in the AZ group which suggests that either DMARD therapy or IMID may accelerate the Pfizer vaccine decay rate. Since an mRNA vaccine is preferred for booster doses in adults, these findings suggest that patients with IMID receive a booster dose approximately every 7, 5 and 4 months for those on csDMARD, bDMARD and tsDMARD, respectively.21 This is in support of a recent study showing the risk of severe COVID-19 outcomes is increased after 10 weeks of completing primary vaccination dosing especially in those aged over 80 years, have a greater number of comorbidities, males or those receiving immunosuppressants.22

Our control group data are also in keeping with another study showing that the SARS-CoV-2 antibodies from vaccination waned over time resulting in an approximate 53% and 89% loss within 90 days of vaccination for the Pfizer and AZ vaccinations, respectively.11 In addition, our longitudinal data showed that age had no difference in the SARS-CoV-2 ab titre decay rate.11

Heterogeneity in both the magnitude and decay rates have also been reported to other vaccines such as tetanus and measles where decay rates of 6.2%/year and 0.2%/year, respectively, have been reported, however, these decay rates are substantially lower than that following both AZ and Pfizer vaccination as calculated in our study.23 This could be due to the different mechanisms of action, types of adjuvants used and patterns of loss of protective immunity elicited by the toxoid antigen (tetanus) and live attenuated vaccines (measles) versus viral vector (AZ) and messenger RNA vaccines (Pfizer and Moderna), which can influence the kinetics of antibody production and decay. Of interest, the SARS-CoV-2 Ab decay rate in the csDMARD group was not significantly different to both the AZ and Pfizer control groups. This could be due to the low mean age and dose of methotrexate in our study (54.2 years and 16.9 mg weekly, respectively).

In contrast to already published data which showed that immunogenicity to COVID-19 vaccines may be suboptimal or absent following the second and booster doses due to specific DMARDs such as corticosteroids, mycophenolate and rituximab, our study did not include these agents and hence a direct comparison cannot be made.24 25 In addition, all subjects included in our study were in clinical disease remission and none had interstitial lung disease or renal involvement. Our study, however, observed a significantly diminished humoral response in the tsDMARDs when continued contemporaneous to vaccination.

The methodological limitations to our study have previously been discussed, however, there were additional limitations identified during the 6 months and post booster follow-up.6 First, due to pandemic timelines and the government recommendation to have an earlier booster during the study, the interval from the second to third dose was shorter in some participants.21 This consequently resulted in a number of participants missing their scheduled blood tests. The shortened timeline could also lead to higher residual antibody levels prior to the booster dose and hence under estimate the antibody decay rate. In addition, it also could potentially underestimate the peak antibody response following the booster as the shorter time interval may have an impact on T cells and immunological memory.

Second, the majority of included patients received an mRNA vaccine for the booster dose as this was the preferred vaccine schedule as per the Government guidelines during the study period and hence a homologous schedule for the AZ arm could not assessed.21 A study has observed a 25–35 fold increase in the mean SARS-CoV-2 IgG antibody levels when AZ primed subjects were given a single booster with an mRNA COVID-19 vaccine compared with a three fold increase when AZ booster was used.26 It has also been shown that a homologous Pfizer regime with a prolonged priming interval of 12 weeks instead of 4 weeks had a slightly reduced rate of antibody decay, however, the greater effect on immunogenicity was the choice of vaccine used.27

Third, antigen specific T-cells are also involved following COVID-19 vaccination and relying solely on humoral responses to determine booster timing in patients with and without IMID are not entirely accurate. Hence, reductions in the antibody titres may not translate to a clinically significant loss of protection, however, evaluating IgG levels is still clinically important as convalescent plasma therapy has demonstrated efficacy in critically ill patients with COVID-19 infection.28

Lastly, there is significant heterogeneity in our study population with different IMID’s, DMARDs and COVID-19 vaccines with the assumption that all DMARDs and IMID’s have the same impact on the vaccine response. Our study was, therefore, not adequately powered to examine the vaccine response in each of the IMID and DMARD groups.

Conclusion

Six months after standard vaccination regimens, patients with IMID had protective levels of SARS-CoV-2 IgG Ab similar to controls, except for those on tsDMARDs. Despite higher SARS-CoV-2 IgG Ab decay rates in the Pfizer group, antibody persistence was longer due to a higher peak antibody level following the second vaccination. Seroconversion following the booster dose was 100% across all categories illustrating that a third mRNA-based booster vaccine can restore humoral immunity in patients with IMID on DMARD therapies suggesting that subsequent boosters can elicit robust responses.