Risk of disease flares after SARS-CoV-2 mRNA vaccination in patients with systemic lupus erythematosus

Abstract This study aims to elucidate the effectiveness and safety of SARS-CoV-2 mRNA vaccination in patients with systemic lupus erythematosus (SLE). We enrolled uninfected SLE patients who received two vaccine doses (BNT162b2 or mRNA-1273) and historical unvaccinated patients. Neutralizing antibodies, adverse reactions, and disease flares were evaluated 4 weeks after the second vaccination. Ninety patients were enrolled in each group. Among the vaccinated patients, SLE Disease Activity Index (SLEDAI), and prednisolone doses before vaccination were 2, and 5 mg/d, respectively. After the second vaccination, 19 (21.1%) had no neutralizing antibodies. Adverse reactions occurred in 88.9% within 3 d. Negative antibodies were associated with anemia and mycophenolate mofetil administration. SLEDAI increased modestly but significantly after vaccination, with 13 (14.4%) experiencing flares and 4 (4.4%) severe flares (nephritis in three and vasculitis in one). The flare rate was higher in vaccinated patients than unvaccinated controls. The mean duration between the second vaccination and flares was 35 d, and flares occurred at least 8 days after vaccination. Multivariable analysis showed that high SLEDAI and anti-dsDNA antibodies were associated with flares. The vaccine type, neutralizing antibody titer, and adverse reaction frequency did not affect flares. Therefore, residual disease activity before vaccination increases flare risk.


Introduction
Systemic lupus erythematosus (SLE) is a chronic systemic autoimmune disease characterized by an aberrant immune system and an immunocompromised state due to immunosuppressive treatment.It can be a risk factor for poor outcomes in coronavirus disease 2019 (COVID-19) [1,2].The SARS-CoV-2 mRNA vaccine has significantly reduced the risk of infection and severe COVID-19 clinical events in public health and immunocompromised populations [3].However, patients with SLE need to consider the possibility of underlying disease flares when receiving the vaccine, and this concern may contribute to the fact that patients tend to hesitate to be vaccinated [4,5].Previous reports have shown that flares after vaccine administration range from 3 to 20% in patients with SLE [6][7][8][9][10][11]. Risk factors for flares include a history of recent flares, serological activity, and a history of arthritis and discoid lupus erythematosus [7][8][9].In addition, patients with SLE who receive immune suppressants have lower vaccine efficacy than healthy individuals or those who do not receive these drugs [12][13][14].However, no significant changes were observed in the SLE disease activity index (SLEDAI) before and after vaccine administration.No difference was observed in the frequency of flares compared to unvaccinated individuals during the same observation period [10].Although this information is essential when assessing the risk-benefit balance of mRNA vaccines [10,15], there are still uncertainties.
This study aimed to elucidate the effects of SARS-CoV-2 mRNA vaccination, including adverse reactions to vaccines, neutralizing antibody titers, disease flare rates after vaccination, and associated factors in patients with SLE.

Patients
The study included patients who met the 1999 revised criteria of the American College of Rheumatology (ACR) for the classification of SLE or the 2010 Systemic Lupus International Collaborating Clinics (SLICC) classification criteria [16,17].Consecutive patients who received two vaccine doses (BNT162b2 or mRNA-1273) were recruited between June and October 2021.As a historical unvaccinated control group, other consecutive patients who had not received SARS-CoV-2 vaccination were also recruited.The control patients were selected among those who were in maintenance phase at least 6 months after induction therapy and attended between June and October 2020 during the same season a year earlier with the vaccinated group.They were matched by age and SLEDAI with the vaccination group.The study was approved by the ethics committee of the Keio University School of Medicine (No. 2014-0093) and was conducted in accordance with the Declaration of Helsinki and Good Clinical Practice.All participants provided written informed consent.

Clinical assessments of flares and safety profiles
Demographic and clinical characteristics were collected by reviewing the medical charts.Patients were observed from baseline before the first vaccine dose until the next visit after the second vaccine dose.The components of the Lupus low disease activity state (LLDAS) and SLE Disease Activity Index 2000 (SLEDAI-2K) were assessed at every visit [18].The interval between follow-up visits was at the discretion of the attending physician but mostly every 1-2 months in vaccinated patients and every 1-6 months in the control group.Disease flares based on the Safety of Estrogens in Lupus Erythematosus National Assessment-SLE Disease Activity Index (SELENA-SLEDAI) [19] were evaluated until the next visit after the second vaccine dose in vaccinated patients and until the next visit after the baseline visit in the same season in the control group.Safety profiles were collected from questionnaires regarding adverse reactions after vaccine administration until 4 weeks after the second dose of the vaccine.

Measurement of the neutralizing antibodies
Serum samples were collected from the participants approximately 4 weeks after the second dose of the vaccine.Serum-neutralizing antibodies were measured using the STACIA Analyzer (LSI Medience Corporation, Tokyo, Japan), which measures the inhibitory activities of the patients' anti-SARS-CoV-2 antibodies on the interaction between the receptor-binding domains of the viral spike proteins and human angiotensin-converting enzyme 2 with the STACIA SARS-CoV-2 Neutralization Antibody Test (MBL Corporation, Nagoya, Japan), according to the manufacturer's instructions [20].The cut-off level of seroconversion was defined as an antibody concentration of >1.67 U/mL according to the titers of the 99 percentile of healthy controls.To evaluate the asymptomatic or undiagnosed COVID-19 infection during the study period, we used the HISCL SARS-CoV-2 nucleocapsid (N)-IgG reagent (Sysmex Corporation) according to the manufacturer's instructions.

Statistical analysis
Continuous data are presented as medians and interquartile ranges (IQR), mean and standard deviation (SD), or numbers with percentage values, as appropriate.Wilcoxon's rank sum test and Wilcoxon's signed-rank test were used to examine the differences between continuous variables.Logistic regression analysis and Fisher's exact test were used to compare continuous and categorical data between groups.Multivariable logistic regression was used to identify factors relevant to disease flares.Statistical significance was set at p < .05.All statistical analyses were performed using JMP Pro 17 software (SAS Institute Inc., Cary, NC, USA).

Seroconversions and adverse reactions after the second dose of vaccine administrations
Neutralizing antibodies were measured at 27.6 ± 9.4 days as the mean and SD after the second dose of the vaccine.The median antibody titer was 4.9 (IQR: 1.8-14.9)U/mL.Nineteen (21.1%) patients tested negative after the second dose of the vaccine.Sixty-four (88.9%) patients experienced adverse reactions within 3 d (mean ± SD: 0.7 ± 0.7 d) after the second dose of the vaccine (Supplementary Table 2), including local reactions in 73.6% and systemic reactions in 79.2%.The most prevalent manifestation was local pain in 52 (72.2%).The mean duration of adverse reactions was 2.2 ± 1.6 d.

Disease activity and flares of SLE after the second dose of vaccine administrations
Regarding disease activity, the SLEDAI modestly but significantly increased after the second dose of the vaccine (p = .016,median change 0 [IQR: 0-1]) (Figure 1(A)).Thirteen (14.4%) and four (4.4%) patients experienced flares and severe flares (nephritis in three and vasculitis in one), respectively (Figure 1(B), Supplementary Table 3).The mean duration from the second dose of the vaccine to flares was 35.0 ± 21.4 d, and all flares occurred at least 8 d after the vaccine.Among the patients with flares, nine (69.2%) began additional treatments, including an increase and addition of glucocorticoids, immune suppressants, and biologics in eight and topical intervention with an infusion of vasodilator in one.All flares in the remaining four (30.8%) patients without treatment intensification were mild/ moderate, and their manifestation persisted for an average of 120.8 ± 83.5 d.
Among the 13 patients who experienced flares after the second dose of the vaccine, two patients terminated additional vaccination at that time.Among the remaining 11 patients who received the third vaccine, one patient who had experienced a severer flare of nephritis with worsening of proteinuria, haematuria and leukocytopenia after the second vaccination underwent a mild/moderate flare of an increase in SLEDAI score of 4 points due to an increase in proteinuria.2).Multivariable logistic regression analysis with two models, in which SLEDAI and anti-dsDNA antibodies were separated owing to multicollinearity, showed that high levels of SLEDAI (p = .031,OR 1.301 [IQR: 1.014-1.670])and anti-dsDNA antibodies (p = .048,OR 1.021 [IQR 1.000-1.042])were associated with flares.The proportion of azathioprine administration was significantly higher in the flare-up group only in multivariable model 1 with SLEDAI and not in model 2 with anti-dsDNA antibodies.Flares were unrelated to HCQ use, vaccine type, neutralizing antibody titers, and adverse reactions or flares in the year before the first dose of vaccine.

Associated factors with flares of SLE after the second dose of vaccine administrations
In one patient who experienced flares after both the second and third vaccinations, the flare after the third vaccination was milder than that after the second vaccination, as the proteinuria improved without additional treatment and there were no changes in anti-dsDNA antibodies or serum complement levels.The patient had received remission induction therapy at the time of flare after the second vaccination and had improved, and at the time of the third vaccination, anti-dsDNA antibodies were negative and the SLEDAI score was 0. This course of events supported the above results of associated factors with flares after the second vaccination.

Comparison in the vaccine-administered patients and the historical unvaccinated controls
Finally, we compared the patients with second dose of the vaccine (n = 90) with the historical unvaccinated control patients (n = 90).The baseline clinical characteristics did not differ between the two groups except for the dose of PSL (p = .012,median 5.0 [IQR: 3-8] mg/d vs 4.3 [0-6] mg/d) (Supplementary Table 1).In the unvaccinated control group, four (4.4%) patients experienced a flare and one (1.1%)patient experienced flares and a severe flare (change in SLEDAI score of 6 points with additional MMF) during the mean period of 72.7 ± 32.9 d.The vaccine-administered patients showed a significantly higher flare incidence and slightly increased SLEDAI compared to the unvaccinated controls (p = .039,13 [14.4%]vs 4 [4.4%] and p = .018,median 0 [0-1] vs 0 [0-0], respectively).The changes in anti-dsDNA antibodies were not different (Table 3).

Discussion
This study indicated that the disease activity modestly but significantly increased after the second dose of the SARS-CoV-2 mRNA vaccine in patients with SLE.Patients with residual disease activity and serological activity before vaccine administration may have a higher risk of flares.However, no association was observed between the flares, seroconversion or adverse reactions.
The proportions of negative neutralizing antibodies and disease flares were comparable with those reported previously [6,8].Among the extracted factors associated with negative neutralizing antibodies in this study, MMF administration supported previous results.Administration of prednisolone and methotrexate and low IgG levels were not associated with them, and only one patient received belimumab and none received anifrolumab nor rituximab in this cohort.In the previous reports in African infants, anemia and transferrin receptors have been identified as risk factors for non-seroconversion to vaccines against diphtheria and pneumococcus [21].The mechanism is considered that iron deficiency is linked to the impairment of the acquired immune system.Although iron deficiency may potentially underlie the connection between anemia and seroconversion in this cohort, comprehensive data elucidating the association with seroconversion are lacking, as anemia can be influenced by secondary factors such as chronic inflammation and medication use in patients with SLE.Disease flares and adverse reactions to vaccines were distinguishable in the present cohort.This is because the mean onset of flares confirmed was 27.8 d after the second dose of the vaccine, with a persistent duration of >1 month and 69.2% requiring some intervention.In contrast, adverse reactions occurred within 3 d and resolved within 7 d.
Previous studies reported no disparity in SLEDAI changes between patients receiving the second dose of the vaccine and unvaccinated individuals [10].In contrast to those studies, our current findings present that vaccine-administered patients with SLE exhibited a higher flare rate and a modest increase in SLEDAI compared to the unvaccinated controls.However, the incidence of severe flares showed no significant difference between the two groups, and  .520 Bold numbers indicate that the difference was significant, with a p-value <.05.Fisher's exact test was used to analyse categorical variables, and logistic regression analysis was used for continuous variables.SledaI and anti-dsdna antibodies and rash were divided for 2 models in multivariable analysis because of multicollinearity.aZ: azathioprine; cI: calcineurin inhibitor; egFR: estimated glomerular filtration rate; HcQ: hydroxychloroquine; mmF: mycophenolate mofetil; mTX: methotrexate; PSl: prednisolone; SledaI: systemic lupus erythematosus disease activity index; uPcR: urine protein creatinine ratio.
alterations in serological parameters did not align consistently with SLEDAI.The discrepancies in the results between our results and the previous report may be attributed to the difference in patient backgrounds.Notably, unvaccinated controls in the present study were selected from a different group of patients in the same season a year earlier.In the vaccine-administered patients, flares during the year before the first vaccine were observed in 20% of the patients, whereas flares within 1 month after the second dose were observed in 14.4% of the patients.
Although it was impossible to compare the two directly because of their different backgrounds, no relationship was found between flares in the past year and flares after vaccine administration.SLE disease activity may increase after vaccine administration due to the cross-reactivity of antibodies to in vivo proteins that share common parts with the SARS-CoV-2 spike protein as molecular mimicry [22,23], toll-like receptor (TLR) stimulation and increased interferon-α production from plasmacytoid dendritic cells induced by mRNA vaccines, and induction of pathogenic B cells in susceptible patients [24][25][26].Previous reports of more subsequent flares of SLE in the higher interferon-α group at the time of influenza vaccine administration suggested that the associations between residual disease activities and flares were not specific to mRNA vaccines [24].However, a meta-analysis reported that influenza and pneumococcal vaccines do not impact disease activity scores in patients with SLE [27].The quadrivalent human papillomavirus vaccine also had the same frequency of increased disease activity as the matched SLE control [28].However, autoreactive T cells in the peripheral blood in repeated SARS-CoV-2 mRNA vaccine-administered patients with SLE have decreased [29], and nucleoside modification of mRNA vaccines prevents the stimulations of TLRs and reduces interferon-α signaling [30].Therefore, SARS-CoV-2 mRNA vaccine administration may protect against some autoimmune phenomena, and the balance between disease promotion and suppression may be antagonistic.Therefore, it is conceivable that the susceptible population in the present cohort may have flared up with residual disease and serological activity, whereas it may have remained mild.
A limitation of this study is that the cohort was limited to a small number of single-center Japanese patients; no validation was observed after the third dose of vaccine administration, the background immunological evaluation may not have been sufficient, and the existence of confounding factors that could not be analyzed and validated in this study cannot be ruled out.A study reported that third SARS-CoV-2 vaccination induced enhanced humoral and cellular immunogenicity in patients with SLE, without any indications of increased disease activity [31].The impact of a fourth vaccination on the Omicron variant remains uncertain, and cases of SLE flares after fourth doses of vaccination were observed in previous reports [32].To the best of our knowledge, information regarding the efficacy and safety of vaccine administration beyond the fourth dose in patients with SLE was unavailable.Furthermore, the correlation of cellular and humoral immune response in SLE patients was reported to be inoculated with BNT162b2 [6,15].We could not assess the humoral response to more than 3 doses and cellular response to COVID-19, which has been suggested to play an important role [33], and it is a major limitation of this study.However, we believe that assessing the humoral response to two vaccines and the relationship between disease flare and vaccinations would provide useful information to establish the strategy of vaccinations in patients with SLE.
The results of this study with mRNA vaccines support the recommendation for vaccination by the European League against Rheumatism that administration should preferably be in stable disease and individually [34].The risk-benefit balance of the mRNA vaccine in patients with SLE is favorable; however, careful follow-up is required if there is residual disease activity of critical organ damage, and further studies are needed to clarify if other vaccines have the same effect.
Our findings allowed us to consider the risk-benefit balance of SARS-CoV-2 mRNA vaccine administration in patients with SLE.While a neutralizing antibody-negative population has been identified, residual disease activity and serological activity before vaccine administration can be risk factors for flares after vaccination.Careful follow-up is required after vaccine administration in patients susceptible to residual disease activity, particularly in critical organs.

Figure 1 .
Figure 1.change in systemic lupus erythematosus disease activity index (SledaI) and flare rates at the subsequent visit after the second dose of the vaccine.changes in SledaI from baseline to the subsequent visit after the second dose of the vaccine (a) and flares based on the Selena-SledaI flare index (B) were evaluated.The mean time between the second dose of the vaccine and the subsequent visit to assess disease activity and flares was 29.2 ± 10.9 d.Wilcoxon's rank sum test was used for the analysis.*p < .05.

Table 1 .
Baseline demographics and clinical characteristics in patients with systemic lupus erythematosus and factors associated with non-seroconversion after the second dose of the vaccine.

Table 2 .
Factors associated with disease flares after the second dose of the vaccine.

Table 3 .
comparison in flares, disease activity and serological parameters between the vaccine-administered patients and the unvaccinated controls.asmedian (IQR) unless otherwise specified.Bold numbers indicate that the difference was significant, with a p-value <.05.Fisher's exact test was used to analyse categorical variables.Wilcoxon's rank sum test was used for continuous variables between the groups, and Wilcoxon's signed-rank test was used for continuous variables comparing changes in the respective group.The change in the parameters indicated by Δ was calculated by subtracting the baseline visit from the subsequent visit.