Analysis of vaccine adverse events associated with rheumatoid arthritis: A pharmacovigilance study based on the VAERS database Authors: Pan Lu, Mengqi Lou, Zhihao Qiu, Ruibo Xia, Jialu Shu, Guifang Lu Published: 2026 ABSTRACT ABSTRACTVaccination is a principal method of infectious disease prevention, but its association with rheumatoid arthritis (RA) remains controversial. This study assessed vaccine‑associated RA using the Vaccine Adverse Event Reporting System (VAERS). Data from 1990 to 2026 were extracted. Descriptive statistics, Weibull fitting, and disproportionality analysis using four methods, including the reporting odds ratio (ROR), along with subgroup analyses by age, sex, and pre‑/post‑COVID, were performed. Among 11,532,185 reports, 34,498 RA‑related adverse events (AEs) (0.03%) involved 5006 subjects. Females comprised 77.9% and 18–65 predominated. Serious outcomes occurred in 13.48%. Most AEs (47.49%) were musculoskeletal. 47.0% occurred within 7d (median 5.4), indicating early failure. The COVID‑19 vaccine had the most reports (n = 24,266) but a weak signal (ROR = 1.25); the Lyme disease vaccine (ROR = 27.81), the rubella vaccine (ROR = 5.69), and the anthrax vaccine (ROR = 3.90) exhibited the strongest signals. Subgroup signals: Lyme disease vaccine except 2021–2026; rubella vaccine only in females/≤17; anthrax vaccine in 18–64. RA‑related vaccine AEs are extremely rare, mainly musculoskeletal and within 1 week. High COVID‑19 volume did not align with its weak signal, while strong signals for the Lyme disease, rubella, and anthrax vaccines reflected specific populations. No strong disproportionality signal was found. However, VAERS is passive, cannot establish causality or confirm safety.KEYWORDS: Vaccine Adverse Event Reporting System, pharmacovigilance, rheumatoid arthritis, adverse events, signal detection, COVID-19 vaccines Introduction Vaccination is one of the most effective and cost-efficient public health measures for preventing the spread of infectious diseases, reducing severe complications, and lowering mortality. For example, vaccination played an irreplaceable role in the response to the recent COVID-19 pandemic.1 For patients with autoimmune diseases, on the one hand, vaccination is a particularly important preventive measure, as the dual effects of the underlying disease and immunosuppressive therapy markedly increase the risk of infections.2 On the other hand, impaired immune function may reduce vaccine immunogenicity and protective efficacy,3 and some studies have reported that vaccines are more likely to trigger autoimmune events in this population.4 With the widespread uptake of vaccines, public and medical concerns, particularly in regard to special populations, warrant further investigation into vaccine safety. Rheumatoid arthritis (RA) is a common autoimmune disease characterized by symmetric polyarthritis, synovial hyperplasia, and joint destruction. Its complex pathogenesis results from a combination of genetic predisposition, environmental triggers, and subsequent aberrant activation of multiple immune cell types and inflammatory cytokines.5 The prevalence of RA in females is approximately 2.5 times that in males, and the overall prevalence has been rising globally over the past three decades. With an estimated 17 million individuals affected in 2020, the disease burden continues to increase,6 resulting in a corresponding increase in the demand for vaccination among this population. However, many patients with RA receive immunosuppressive therapies, such as methotrexate, tumor necrosis factor (TNF) inhibitors, and other disease-modifying antirheumatic drugs (DMARDs), which may reduce the immunogenicity and protective efficacy of vaccines. This creates a clinical dilemma: although these patients face an elevated risk of infections that could be prevented by vaccines, their underlying immunosuppression may attenuate the vaccine immune response. Existing studies have shown that methotrexate can attenuate antibody responses to multiple vaccines including influenza and COVID-19 vaccines, while rituximab has a more pronounced and sustained impact, and TNF inhibitors have a relatively milder effect.7 These findings suggest that vaccination strategies for RA patients need to be individualized according to their immunosuppressive regimen. In recent years, a number of clinical case reports and case series have suggested that vaccination may be followed by new-onset RA.8 For patients with existing RA, vaccination may exhibit poor efficacy9 or be followed by serious adverse events (AEs).10 In contrast, a recent meta-analysis-involving nearly 26 million participants found no clear association between vaccination and RA risk.11 However, these studies mainly focused on common vaccines, such as COVID-19-, influenza, and HPV; less commonly used vaccines are more difficult to evaluate in clinical trials. Moreover, vaccine trials in special RA populations, such as elderly, pregnant, or pediatric patients, have strict ethical limitations. Therefore, the relationship between vaccines and RA lacks systematic conclusions based on real-world data-, further exacerbating vaccine hesitancy among the RA population.12 The Vaccine Adverse Event Reporting System (VAERS) is a national vaccine safety surveillance system jointly established by the US Centers for Disease Control and Prevention (CDC) and the Food and Drug Administration (FDA), responsible for collecting and analyzing AEs following immunization.13 As a passive surveillance system, VAERS rapidly captures potential vaccine safety signals, particularly those associated with rare or novel AEs, to provide valuable data for in-depth research and risk assessments.14 To the best of our knowledge, previous studies have conducted data mining in VAERS for systemic lupus erythematosus15 and spondyloarthritis,16 but not specifically for RA. Therefore, this study aimed to systematically evaluate vaccine-associated RA using data from the VAERS database, thus providing evidence-based support for vaccine selection and risk assessment in clinical practice. Methods Data source and acquisition Data for this study were obtained from the publicly available VAERS database (https://vaers.hhs.gov/data/datasets.html), covering the period from January 1990 to March 2026. The data comprised the three datasets available on the VAERS website: VAERS DATA, VAERS Symptoms, and VAERS Vaccine. These datasets include patient demographics, past and current medical history, laboratory data, and AEs. Based on searches by other studies,17 preferred terms (PTs) were searched for using the keyword “Rheumatoid,” based on the Medical Dictionary for Regulatory Activities (MedDRA) version 26.0. After removing duplicate PTs and those unrelated to RA, the following PTs were obtained: arthritis rheumatoid; arthritis rheumatoid aggravated; juvenile rheumatoid arthritis; laryngeal rheumatoid arthritis; malignant rheumatoid arthritis; monoarthritic rheumatoid arthritis; monoarticular juvenile rheumatoid arthritis; other rheumatoid arthritis with visceral or systemic involvement; palindromic rheumatoid arthritis; pauciarticular juvenile rheumatoid arthritis; polyarticular juvenile rheumatoid arthritis, acute; polyarticular juvenile rheumatoid arthritis, chronic or unspecified; progression of rheumatoid arthritis; pulmonary rheumatoid nodule; rheumatoid arthritis; rheumatoid arthritis aggravated; rheumatoid arthritis and associated conditions; rheumatoid arthritis and other inflammatory polyarthropathies; rheumatoid arthritis flare up; rheumatoid arthritis in remission; rheumatoid arthritis relapse; rheumatoid arthritis-associated interstitial lung disease; rheumatoid arthropathies; rheumatoid bursitis; rheumatoid lung; rheumatoid meningitis; rheumatoid neutrophilic dermatosis; rheumatoid nodule; rheumatoid pleuritis; rheumatoid pneumoconiosis; rheumatoid scleritis; rheumatoid spondylitis; rheumatoid vasculitis; seronegative rheumatoid arthritis; seropositive rheumatoid arthritis; spinal rheumatoid arthritis; and systemic rheumatoid arthritis. The downloaded data were cleaned and filtered based on these PTs. Of note, the extracted PTs included terms indicating disease flares or relapses (e.g., ‘rheumatoid arthritis flare up,’ ‘rheumatoid arthritis relapse,’ ‘rheumatoid arthritis aggravated’), which were collectively classified under RA-related AEs. Descriptive analysis RA-related AEs were evaluated in terms of the number of events, sex and age distribution, serious outcomes, and whether only one type of vaccine was received when the RA-related AE occurred. The distribution of all RA-related AEs by system organ class (SOC) was also explored. Time-to-onset analysis To characterize the temporal patterns associated with RA-related AEs, a time-to-onset analysis was performed. The overall time frame was divided into multiple intervals, and AEs were allocated to intervals based on the time between vaccination and the AE report. The number and proportion of events in each interval was counted to characterize the temporal concentration of AEs. A Weibull distribution18 was further fitted to the time intervals to estimate distribution parameters and quantify the time-varying risk characteristics of the AE occurrences. Disproportionality analysis The association between vaccines and RA was assessed using a disproportionality analysis,19 an important tool for interpreting spontaneous AE reports in pharmacovigilance. Four methods were used: reporting odds ratio (ROR), proportional reporting ratio (PRR), Bayesian confidence propagation neural network (BCPNN), and multi-item gamma Poisson shrinker (MGPS). The ROR and PRR were calculated with 95% confidence intervals (CIs). An association was considered significant when the lower bound of the 95% CI exceeded one for the ROR, or when the following threshold criteria were met: PRR ≥2, χ2 ≥4, and ≥3 cases. The BCPNN generated the information component (IC), with IC025 >0 indicating statistical significance. Similarly, the MGPS produced empirical Bayesian geometric mean (EBGM) values, with EB05 >1 considered a positive signal. These complementary approaches provided robust signal detection while mitigating the limitations inherent to the individual methods (e.g., the susceptibility to confounding of ROR, the reduced sensitivity for rare events of PRR, and the computational demands of Bayesian methods). The specificity was enhanced by requiring positive signal concordance in at least two algorithms. Heterogeneity analysis To assess the robustness of the signal detection results and control for potential confounders and reporting bias, subgroup-stratified sensitivity analyses were performed. Considering that demographic characteristics of RA may influence the association between vaccine exposure and AEs, the data were stratified by sex (male, female)20 and age group (≤17 y, 18–64 y, and ≥65 y).21 Additionally, given that the surge in vaccination during the COVID-19 pandemic may have affected the reporting distribution and signal stability,22 the reporting period was divided into two phases: 1990–2020 and 2021–2026. The cutoff was chosen as 2021 because COVID-19 vaccines became widely available in the United States at that time. In each subgroup, in addition to calculating the ROR, PRR, EBGM, and IC with their corresponding 95% CIs or lower bounds, the number and proportion of serious outcomes were reported. This allowed the distribution characteristics of signal strength and clinical severity to be comprehensively evaluated. A signal was considered robust if it persisted across all subgroups with consistent directional indicators and the proportion of serious outcomes did not show significant bias. Conversely, if a signal was confined to specific subgroups or showed obvious population selectivity in the distribution of serious outcomes, it was interpreted with caution. Statistical analysis All statistical analyses were performed using Python 3.14.3 and Excel 2021. Results Clinical characteristics of vaccine-associated RA As of March 6, 2026, VAERS contained 11,532,185 total AE reports, of which 34,498 were RA-related AEs (0.03%) involving 5006 patients. Of the patients in which the sex and age were known, there were significantly more females than males with RA-related AEs, and the predominant age group was 18–65 y. The proportion of serious outcomes was relatively small. Most of the patients, 92.27%, had received only one type of vaccine when the RA-related AE occurred. Detailed data are shown in Table 1. Table 1.Clinical characteristics of vaccine-associated rheumatoid arthritis VAERS reports.CharacteristicsParticipants (N%)Overall5006Female1588 (31.72%)Male449 (8.97%)Unknown2969 (59.31%)Age (years) <18106 (2.12%)18–651111 (22.19%)>65477 (9.53%)Unknown3312 (66.16%)Serious outcomes No4331 (86.52%)Yes675 (13.48%)Life-threatening51 (1.02%)Emergency room visit269 (5.37%)Hospitalization170 (3.40%)Disabled357 (7.31%)One type of vaccine administered Yes4619 (92.27%)No387 (7.73%) Clinical characteristics of vaccine-associated rheumatoid arthritis VAERS reports. Distribution of RA-related AEs by SOC The 34,498 AEs involved 27 SOCs. After excluding four SOCs that were not closely related to clinical manifestations (“investigations,” “surgical and medical procedures,” “social circumstances,” and “product issues”), 22,749 RA-related AEs remained. “Musculoskeletal and connective tissue disorders” accounted for the largest proportion of the AEs (47.49%, n = 10,803), with the main PTs including “rheumatoid arthritis,” “arthralgia,” “pain in extremity,” “joint swelling,” and “myalgia,” as shown in Figure 1. Distribution of vaccine-related adverse events in rheumatoid arthritis by system organ class (SOC). Time-to-onset analysis of vaccine-associated RA The time between the vaccination and RA-related AE reports was analyzed after stratifying the data by time interval (Figure 2). The highest proportion of reports occurred within 0–7 d post-vaccination (47.0%, n = 611), with fewer AEs reported in the subsequent time intervals: 8–14 d (9.7%, n = 126), 15–30 d (11.6%, n = 151), 31–60 d (9.0%, n = 117), 61–120 d (7.5%, n = 97), 121–180 d (4.1%, n = 53), 181–360 d (6.5%, n = 84), and >360 d (4.6%, n = 60). The overall distribution showed a high concentration of AE reports shortly after vaccination that gradually decreased over time. The Weibull distribution fit (Figure 3) further validated this trend, with a median onset time of 5.4 d (IQR: 1.0–50.5 d). A full 25% of AEs occurred within 1 d post-vaccination, and 75% occurred within 50.5 d. The time range spanned from 0 to 6,570 d, indicating that some events were reported during long-term follow-up. The fitted scale parameter α was 13.18 (95% CI: 7.35–17.88), which indicated that approximately 63.2% of RA-related AEs occurred within 13.18 d after vaccination. The shape parameter β was 0.41 (95% CI: 0.35–0.46). β <1 indicated that the risk of AE occurrence was highest early after vaccination, and therefore exhibited an “early failure” pattern. Stratification of rheumatoid arthritis-related vaccine adverse events by time interval. Weibull distribution fitting of rheumatoid arthritis-related vaccine adverse events. Signal strength analysis of vaccine-associated RA After excluding unknown entries, 131 particular vaccines were identified as suspected of causing RA-related AEs. A disproportionality analysis was performed on the top 30 vaccines by number of reported cases (Figure 4). The top five vaccines by number of reports were as follows: COVID-19 vaccine (n = 24,266), IRP (polio vaccine; n = 1762), VARZOS (zoster vaccine live; n = 1409), FLUX (influenza vaccine; n = 1122), and LYME (Lyme disease vaccine; n = 1073). In terms of signal strength, LYME (Lyme disease vaccine) had the strongest signal (ROR = 27.81), followed by RUB (rubella vaccine; ROR = 5.69) and ANTH (anthrax vaccine; ROR = 3.90). HEPAB (hepatitis A and B), TTOX (tetanus toxoid), IRP (polio), and IPV2 (inactivated polio vaccine) also exhibited positive signals. Although the COVID-19 vaccine had the highest number of reports, its signal was weak (ROR = 1.25). VARZOS (zoster), FLU3 (influenza), and MMR (measles-mumps-rubella) had ROR values less than 1, suggesting that their reporting rates were below the background level. Forest plot of the signal strength of the top 30 vaccines. Subgroup analyses by age, sex, and pre- vs. post-COVID-19 period Based on the disproportionality analysis results, potential reporting bias was further evaluated by conducting subgroup analyses of the three vaccines with the highest ROR values (Lyme disease vaccine, rubella vaccine, anthrax vaccine) by sex, age group, and pre- vs. post-COVID-19 period, and by examining the signal strength and proportion of serious outcomes (Table 2). The three vaccines exhibited marked population heterogeneity and temporal characteristics. The Lyme disease vaccine signals were extremely strong and stable across all sex and age subgroups (ROR of 30.18–246.80, EBGM >27, and IC lower bound >4.48), with serious outcome proportions of 63.16%–77.78%. However, because the vaccine was discontinued in 2002, reports were not available after 2020; therefore, the signal was only evaluated from 1990–2020 (ROR = 36.25). The RUB vaccine signal exhibited significant sex and age bias. It was statistically associated only in females (ROR = 5.68) and children ≤17 y old (ROR = 45.69). No signal was found in males or individuals ≥65 y old, which suggested that the signal was primarily driven by specific vaccinated populations. The anthrax vaccine signal was almost exclusively present in adults aged 18–64 y old (males: ROR = 5.62, serious outcome proportion 88.89%; females: ROR = 7.49), with no signal in those ≤17 or ≥65 y old. However, the signal disappeared in the second evaluation period (2021–2026; ROR = 0.59, 95% CI lower bound 0.24). These findings were consistent with the target population of the vaccine (primarily military personnel aged 18–64 y old) and the trend toward reduced vaccination in recent years. In summary, the Lyme disease vaccine signal was extremely robust among historically vaccinated populations, the rubella vaccine signal exhibited clear population bias, and the anthrax vaccine signal was strictly population-specific with recent attenuation. Table 2.Signal strength and serious outcomes of LYME (lyme disease), RUB (rubella), and ANTH (anthrax) vaccines stratified by age, sex, and pre- vs. post-COVID-19 period.VaccinetypeSubgroupROR (95% CI)PRR (χ2)EBGM (EBGM05)IC (IC025)Number of serious outcomes (%)LYMEFemale30.18 (27.46–33.18)28.16 (12083.82)27.09 (24.64)4.76 (4.54)24 (63.16%)Male58.29 (53.25–63.80)53.53 (26189.50)46.08 (42.10)5.53 (5.29)33 (70.21%)≤17246.80 (172.31–353.49)212.76 (7238.78)202.89 (141.65)7.66 (4.46)1 (50.00%)18–6431.13 (28.80–33.65)29.08 (18439.75)26.99 (24.97)4.75 (4.59)45 (71.43%)≥6543.17 (34.88–53.42)39.41 (3476.55)38.46 (31.08)5.27 (4.48)7 (77.78%)1990–202036.25 (33.90–38.75)33.61 (29122.34)28.90 (27.03)4.85 (4.72)57 (66.28%)2021–20260.00 (0.00–0.00)0.00 (0.01)0.00 (0.00)−inf (−3.24)0 (0.00%)RUBFemale5.68 (4.15–7.76)5.61 (151.49)5.60 (4.09)2.48 (1.88)4 (40.00%)Male0.00 (0.00–0.00)0.00 (0.72)0.00 (0.00)−inf (−3.67)0 (0.00%)≤1745.69 (22.53–92.67)44.35 (335.73)43.90 (21.65)5.46 (1.94)0 (0.00%)18–644.01 (2.72–5.90)3.98 (57.97)3.97 (2.70)1.99 (1.28)3 (42.86%)≥650.00 (0.00–0.00)0.00 (0.09)0.00 (0.00)−inf (−3.05)0 (0.00%)1990–20206.81 (5.30–8.74)6.71 (304.08)6.66 (5.18)2.74 (2.25)4 (30.77%)2021–20260.00 (0.00–0.00)0.00 (0.38)0.00 (0.00)−inf (−3.36)0 (0.00%)ANTHFemale7.49 (6.59–8.51)7.37 (1320.09)7.24 (6.37)2.86 (2.63)7 (63.64%)Male5.62 (5.04–6.26)5.57 (1246.20)5.18 (4.65)2.37 (2.20)16 (88.89%)≤170.00 (0.00–0.00)0.00 (0.20)0.00 (0.00)−inf (−3.16)0 (0.00%)18–644.91 (4.51–5.33)4.86 (1701.59)4.63 (4.26)2.21 (2.08)21 (84.00%)≥650.00 (0.00–0.00)0.00 (0.11)0.00 (0.00)−inf (−3.07)0 (0.00%)1990–20204.97 (4.57–5.40)4.92 (1733.56)4.60 (4.23)2.20 (2.07)23 (82.14%)2021–20260.59 (0.24–1.41)0.59 (1.46)0.59 (0.24)−0.77 (−1.85)0 (0.00%) Signal strength and serious outcomes of LYME (lyme disease), RUB (rubella), and ANTH (anthrax) vaccines stratified by age, sex, and pre- vs. post-COVID-19 period. Discussion This study provides a comprehensive pharmacovigilance analysis of vaccine-associated RA based on data from the world’s largest vaccine AE reporting system, covering the period from January 1990 to March 2026. To our knowledge, this study evaluated the largest sample size and the longest time span of RA-related vaccine AE events to generate the most systematic evaluation to date. Our results showed that RA-related AEs were extremely rare in the VAERS database (0.03%). The reported clinical manifestations predominantly involved musculoskeletal and connective tissue disorders, and there was a relatively low proportion of serious outcomes. No strong disproportionality signals were observed among the most widely used vaccines, consistent with the absence of disproportionate safety signals for routine vaccinations in patients with RA in passive surveillance data.7 We observed a strong temporal dependence in reports of RA-related AEs, which peaked during week 1 post-vaccination (47.0%) with a median onset time of 5.4 d. This early failure pattern was consistent with reports on other systemic autoimmune rheumatic diseases,23 suggesting a potential temporal association, though this does not establish a causal relationship with a very short latency period. This rapid onset supports a temporal association between vaccination and RA events, but does not imply causality. The AEs may be related to a robust activation of the innate immune system by vaccine adjuvants,24 which may trigger a local inflammatory cascade in the joints of genetically susceptible individuals. In the signal strength analysis, significant heterogeneity was observed in the associations between different vaccines and RA-related AEs. Notably, the COVID-19 vaccine had the highest number of AE reports (accounting for >70% of the total reports in this analysis), but its signal strength was weak (ROR = 1.25). This suggested that the large number of reports may be primarily due to the extremely high vaccination coverage during the same period rather than an intrinsic risk of the vaccine to trigger RA. This finding is consistent with pharmacovigilance analyses of patients with RA following COVID-19 vaccination, which reported that vaccination was not associated with a significant increase in disease activity or new-onset RA in the study population.25 Specifically, mRNA-based vaccination did not lead to increased autoantibody titers or disease flares, and successive doses were even associated with a modest decrease in rheumatoid factor titers. Similarly, another real-world study confirmed adequate immunogenicity and a low incidence (1.7%) of mild disease flares post-vaccination.26 These converging lines of evidence support that the high volume of AE reports is more likely a reflection of the massive vaccination scale rather than a causal relationship with RA exacerbation. In contrast, specific vaccines such as Lyme disease vaccine, rubella vaccine, and anthrax vaccine showed very strong signals, warranting further evaluation. The Lyme disease vaccine exhibited the strongest signal (ROR = 27.81), but this was based only on data available before 2002, prior to its market withdrawal due to safety concerns.27 Moreover, Lyme disease is primarily transmitted by tick bites, and vaccination alone cannot address the broader public health challenges caused by ticks;28 Therefore, the findings on Lyme disease vaccine and RA-related AEs do not warrant further discussion. The rubella vaccine signal had a clear population bias and was significant only in females and children ≤17 y. This finding is consistent with the vaccine primarily being administered to children,29 and with females generally producing higher antibody responses to vaccines and experiencing more AEs than males.30 An early VAERS data mining study reported that the rubella vaccine was associated with various arthritic reactions,31 but subsequent studies did not find a clear association between the rubella vaccine and arthritis. Furthermore, the measles – mumps – rubella booster vaccination demonstrated a favorable safety profile in patients with juvenile idiopathic arthritis.32 The anthrax vaccine signal was highly concentrated in adults aged 18–64 y old, consistent with its administration primarily being restricted to high-risk occupational groups. A case-control study in a military population showed that anthrax vaccination was associated with recent-onset RA but did not increase the long-term risk of RA.33 However, anthrax vaccine-related AEs have shown high signal values for rheumatic diseases globally.34 Overall, the subgroup analysis characteristics of these three high-signal vaccines were closely related to their respective vaccinated populations and immunization strategies, and these disproportionality analyses did not raise any new safety concerns. However, the above findings do not completely rule out the risk of vaccination in RA patients, as VAERS lacks data on concomitant immunosuppressive medications. Given that immunomodulatory therapies may influence vaccine immunogenicity and subsequent AE profiles, this issue warrants further discussion. An important clinical consideration in patients with RA is the effect of immunosuppressive therapy on vaccine immunogenicity and efficacy. Methotrexate, the most commonly used DMARD in patients with RA, has been shown to attenuate antibody responses to various vaccines, including SARS-CoV-2 vaccines. Similarly, TNF inhibitors and other biologics may disrupt the immune response to vaccination.35 Therefore, vaccination timing should be optimized relative to immunosuppressive therapy cycles. For example, temporarily discontinuing methotrexate following COVID-19 vaccination has been shown to improve vaccine immunogenicity without significantly increasing RA disease activity.36 Furthermore, given that patients with RA have an elevated risk of vaccine-preventable infections, the available evidence suggests that the benefits of vaccination may outweigh the potential risks, and rheumatology guidelines strongly recommend that patients with RA receive routine immunizations.37 Future studies should investigate whether specific immunosuppressive regimens differentially affect the risk of RA-related AEs following vaccination. The limitations of this study are similar to those of other studies based on spontaneous reporting systems and should be interpreted with caution. First, VAERS is a passive surveillance system with inherent reporting biases, including underreporting, delayed reporting, and the Weber effect.38 This was apparent in the present study, in which the number of AEs associated with the COVID-19 vaccine far exceeded those associated with other vaccines, although the signal was weak. Therefore, the association analyses in this study represent statistical disproportionality, not causality. Second, the vaccine data does not include the total doses administered for each vaccine. Therefore, the precise incidence rates cannot be calculated, and the possibility that changes in reporting rates were driven by changes in vaccination volume cannot be excluded. Furthermore, information on vaccine platforms was missing for various vaccines, which precluded an in-depth discussion of the impact of different vaccine platforms on RA-related AEs. The influence of vaccine platforms on vaccine efficacy is significant, as demonstrated during the COVID-19 pandemic,39 and the risks associated with different vaccine platforms may differ, particularly in immunocompromised patients, such as those with RA. Third, the RA-related data did not include the ACR/EULAR classification criteria for confirmed RA, which limited our ability to conduct deeper subgroup analyses, such as distinguishing new-onset RA from RA flares or relapses. In the absence of validated classification criteria, some reported events may represent preexisting RA disease activity rather than vaccine-triggered new-onset cases. Furthermore, key RA-related clinical data, such as serological markers (rheumatoid factor and anti-CCP antibodies), acute-phase reactants, and joint imaging findings, are not included in VAERS, and therefore the clinical authenticity and severity of the reported RA events cannot be verified. Despite these limitations, VAERS, as the largest vaccine safety database, allows potential safety signals to be rapidly identified, including rare events and specific subgroups, providing important indicators for subsequent hypothesis-driven research. Conclusion This pharmacovigilance analysis based on the VAERS database showed that RA-related vaccine AEs are rare, with a reporting rate of less than 0.03%. Their clinical manifestations are predominantly musculoskeletal, and they primarily occur within 1 week post-vaccination. The strength of the association between different vaccines and RA varied significantly: although the COVID-19 vaccine had the highest number of reports, its intrinsic risk signal was weak, whereas the signal strengths of the Lyme disease vaccine, rubella vaccine, and anthrax vaccine were closely related to specific vaccinated populations and immunization strategies. These findings provide pharmacovigilance evidence from a large passive surveillance database. Clinicians and public health decision-makers weighing the benefits and risks of vaccination in patients with RA may consider that the absence of strong disproportionality signals from VAERS for the most commonly used vaccines suggests that routine vaccination may not be associated with a disproportionately increased reporting of RA-related AEs. However, the use of certain vaccines (Lyme disease vaccine, rubella vaccine, anthrax vaccine) in specific populations warrants careful individual risk assessment. As VAERS cannot establish causality or provide incidence rates, these signals should be further investigated in studies with appropriate denominator data and clinical case validation. In future studies, combining electronic health records or registry data with complete clinical information and accurate denominator data may support the further validation and exploration of the potential signals identified in this study.