Association of COVID-19 Vaccinations With Flares of Systemic Rheumatic Disease: A Case-Crossover Study

Abstract

Objective: We aimed to determine the association of COVID-19 vaccination with flares of systemic rheumatic disease (SRD).

Methods: Adults with systemic rheumatic disease (SRD) in a single-center COVID-19 Rheumatology Registry were invited to enroll in a study of flares. COVID-19 vaccine information from March 5, 2021, to September 6, 2022, was obtained from chart review and self-report. Participants self-reported periods of SRD flare and periods without SRD flare. "Hazard periods" were defined as the time before a self-report of flare and "control periods" as the time before a self-report of no flare. The association between flare and COVID-19 vaccination was evaluated during hazard and control periods through univariate conditional logistic regression stratified by participant, using lookback windows of 2, 7, and 14 days.

Results: A total of 434 participants (mean ± SD age 59 ± 13 years, 84.1% female, 81.8% White, 64.5% with inflammatory arthritis, and 27.0% with connective tissue diseases) contributed to both the hazard and control periods and were included in analysis. A total of 1,316 COVID-19 vaccinations were identified (58.5% Pfizer-BioNTech, 39.5% Moderna, and 1.4% Johnson & Johnson); 96.1% of participants received at least one dose and 93.1% at least two doses. There was no association between COVID-19 vaccination and flares in the subsequent 2, 7, or 14 days (odds ratio [OR] 1.46, 95% confidence interval [CI] 0.86-2.46; OR 1.09, 95% CI 0.76-1.55; and OR 0.85, 95% CI 0.64-1.13, respectively). Analyses stratified on sex, age, SRD subtype, and vaccine manufacturer similarly showed no association between vaccination and flare.

Conclusion: COVID-19 vaccination was not associated with flares in this cohort of participants with SRD. These data are reassuring and can inform shared decision-making on COVID-19 immunization.

Figure 1.

Case-Crossover Design The below graphic depicts the case-crossover design employed in this study. Time is represented by the horizontal black lines and moves left to right; the time axes for two theoretical participants (Participants 1 and 2) are represented over the course of the study period (bounded on right and left by vertical lines). Participants report on the activity of their systemic rheumatic disease throughout the study period (smiling face = no disease flare; frowning face = disease flare). Depending on the status of disease at the time of survey submission, a subject either contributes data as a case—i.e., while experiencing a flare—or crosses over to serve as his/her own matched control—i.e., while not experiencing a flare. Hazard periods (red blocks) encompass the windows of time immediately prior to the onset of flares. Control periods (green blocks) are derived from the windows preceding self-reported non-flares. The occurrence of the exposure of interest—i.e., vaccination (syringe)—is evaluated during hazard and control periods within the same subject (exposure windows shaded completely for Participant 1 and with diagonal lines for Participant 2). Each participant represents his/her own unique stratum for analysis, and multiple control periods and hazard periods within the same participant can be compared.