The BNT162b2, mRNA-1273, and Ad26.COV.2.S are the three vaccines that have been instrumental in managing the COVID-19 pandemic in the USA. Despite these vaccines being granted Emergency-Use Authorization by the FDA, safety and efficacy are unclear among autoimmune rheumatic disease patients as these populations were excluded from clinical trials [1,2,3]. Speculations have been rife that vaccines may trigger autoimmunity, contributing to vaccine hesitancy among those with autoimmune rheumatic disorders such as dermatomyositis (DM) [4]. Therefore, the goal was to evaluate the safety and effectiveness of COVID-19 vaccination among DM patients when compared to vaccinated controls using data from a federated database.

TriNetX (Cambridge, MA) is a multicenter research database that was used in this retrospective cohort study. Validated ICD-10 diagnostic codes and CPT codes were utilized to identify vaccinated patients with and without DM. Inclusion of DM patients was based on a diagnosis of ≥ 1 ICD-10 codes at least 1 year apart. 1:1 propensity score matching (PSM) was then utilized to balance the two cohorts by demographics and comorbidities. Lastly, 1-day anaphylaxis along with 30- and 60-day adverse events of special interest (AESI) as defined by the CDC, breakthrough infection (BI), and all-cause hospitalization (ACH) were assessed using adjusted risk ratios and 95% confidence intervals. To protect the patient health information, TriNetX obfuscates aggregate patient counts ≤ 10 to prevent statistical analysis. Further details regarding this methodology are detailed in previous studies and in the supplement [5].

Before PSM, 1,022,471 vaccinated individuals made up non-DM controls and were compared to 6104 vaccinated DM patients. On average, vaccinated DM patients were composed of older and comorbid patients with higher female and Black representation (Table 1). After PSM, two balanced cohorts of 6103 patients were compared to each another. DM patients did not have a difference in risk for immediate anaphylaxis at 1-day post-immunization (RR: 1.8 (CI: 0.96–3.38), p = 0.06), while absolute risk was minimal for DM patients (0.4%). At 30 days post-vaccination, vaccinated DM patients did not experience a difference in risk for AESI, BI, or ACH compared to the control population. However, at 60 days post-vaccination, the DM group had a greater risk for AESI compared to controls (RR: 1.96 (CI: 1.06–3.61), p = 0.028) with a small absolute risk of 0.6%. No differences in risk for BI and ACH were observed 60 days post-vaccination (Table 2). DMARD and glucocorticoid use did not impact AESI, BI, or ACH at any time interval. Among the three administered vaccines, BI was greater among BNT162b2 (0.9% vs. 0.5%, p-0.026) though this was also the most common vaccine administered (4411/6104).

Table 1 Baseline characteristics of the dermatomyositis and non-inflammatory myositis patient cohorts before and after propensity matching
Full size table
Table 2 Post-vaccination 1-day, 30-day, and 60-day outcomes in patients with a diagnosis of dermatomyositis compared to non-inflammatory myositis controls before and after propensity matching for baseline characteristics. Data is listed as percentage (number) and relative risk (95% confidence interval)
Full size table

Widespread efforts to vaccinate the public in attempts to achieve herd immunity makes it increasingly vital for rheumatologists to have an evidence base by which to address patient queries and avoid misinformation, especially surrounding the issue of immunosuppression [6, 7]. We observed a small absolute risk in DM patients; however, the statistically significant rise in 60-day vaccine adverse events may be attributed to autoimmunity being triggered or change in immunosuppressive treatment in patients with autoimmune diseases, a finding that has also been suggested by previous reports [8]. Nevertheless, the benefits of getting vaccinated greatly outweigh the risks in this population especially given very small absolute percent risk [9]. Limitations of this study include potential errors in coding entry and an inability to determine DM severity at time of vaccination. We hope that future studies like the ongoing COVAD study address this gap [10].

While DM patients experienced higher adverse events compared to matched non-DM patients, these findings should not be a deterrent against vaccination in most cases as overall risk for ACH, BI, or immediate anaphylaxis was not increased. Temporal trends of rising AESI at 60 days call for further studies to explore long-term impacts of vaccination in DM patients, especially given high thrombogenic risk associated with DM. Moreover, increasing trends of all-cause hospitalization in DM patients is a concern and additional data is needed with a specific focus on flares of underlying DM.