2.1 Patients

In this prospective cohort study, IIM patients under the care of the Department of Clinical Immunology, University of Debrecen, Hungary were interviewed at the end of the third wave of the COVID-19 pandemic (May–June 2021). The vaccination of the patients started March 1, 2021. At the time of survey every patient had been given the opportunity to receive the vaccines. The vaccination rate, the type of administered vaccines, and the adverse reactions experienced were assessed by questionnaires. These patients were followed prospectively at the clinic or in outpatient departments until July 2022. A loss of follow-up was avoided by conducting telephone interviews with all patients at the end of the study. IIM disease activity, comorbidities, COVID-19 infection data, later vaccination data (booster vaccination started in August 2021), immunosuppressive therapy, and laboratory results were also recorded.

All patients had a definite or probable diagnosis of IIM (muscle weakness, high muscle enzyme levels, plus abnormal electromyogram (EMG) with myopathic motor unit potentials, fibrillation potentials, positive sharp waves, increased insertional activity, and/or muscle biopsy features of inflammatory infiltration in polymyositis (PM) or typical skin symptoms in DM) according to Bohan and Peter,¹⁹ or probable/definite IIM according to the EULAR/American College of Rheumatology's (ACR) myositis criteria.²⁰ Patients with inclusion body myositis were excluded. Interstitial lung disease (ILD) was defined as present by radiographic findings (high-resolution computed tomography [HRCT]) and pulmonary function tests (spirometry, diffusing capacity for carbon monoxide). Dysphagia was diagnosed by barium radiography of the esophagus. Informed consent was obtained from the subjects. This study was carried out in compliance with the Declaration of Helsinki. The study was approved by the Institutional Review Board of the University of Debrecen (Ethical permission number: DE RKEB/IKEB-5723-2021).

2.2 Evaluation of myositis disease status

Disease activity was evaluated using the International Myositis Assessment and Clinical Studies Group (IMACS) core set measures (CSM)²¹: the physician global activity visual analogue scale (VAS), the Manual Muscle Test (MMT-8), patient global activity VAS, a health assessment questionnaire (HAQ), laboratory values of muscle enzymes (CK, creatine kinase; LDH, lactate dehydrogenase), and Myositis Disease Activity Assessment Tools (MDAAT) were assessed at each visit. Active disease at/prior vaccination was defined by physician global activity ≥2 cm of VAS.

Myositis relapse was defined as any of the following: (1) physician-assessed global worsening by ≥2 cm on a 10-cm VAS and worsening on MMT-8 by ≥20%; OR (2) extramuscular organ disease activity worsening by ≥2 cm on a 10-cm VAS; OR (3) any three of six IMACS CSM worsening by ≥30%.^{21, 22}

The criteria for improvement used the six IMACS CSM, combining the absolute percentage change in each with varying weights to obtain a total improvement score (TIS) on a scale of 0–100 using a Web-calculator.²³ Different thresholds of improvement were set for minimal (20–39 points), moderate (40–59 points) and major responses (≥60 points).²⁴

The severity of myositis relapse was graded as minimal, moderate and major. This was determined by a total deterioration score (TDS), which was calculated using the opposite method of the TIS calculation, that is, the absolute percentage worsening change in each IMACS CSM was scored and the sum of the six scores determined the TDS. The relapse category was minimal, moderate, or major based on TDS 20–39, 40–59, or ≥60, respectively.

2.3 Laboratory tests

Routine laboratory markers such as erythrocyte sedimentation rate, total blood count, and chemistry including renal and liver function, ion levels and muscle enzymes (CK, LDH) were tested. Immunological analyses included tests for the following autoantibodies: antinuclear antibodies (ANA), anti-centromere antibodies (ACA), anti-histone antibodies, and anti-cytoplasmic antibodies were determined by indirect immunofluorescence on HEp-2 cells (Viro-Immun Labor-Diagnostika GmbH, Oberursel, Germany); ANA positivity was assessed at 1:40 dilution. Titers of the antibodies against extractable nuclear antigen (ENA) complex, anti-SS-A (Ro) and anti-Jo-1 antibodies, were measured (HYCOR Biomedical Inc., CA, USA) using this latter method. Myositis specific antibodies were detected by membrane-fixed line blots according to the manufacturer's instructions (Euroline Myositis Antigen Profile4, EuroImmun, Lübeck, Germany). Anti-HMGCR antibodies were detected by indirect immunofluorescence assay and confirmed by in-house enzyme-linked immunosorbent assay (ELISA; the cutoff value was ≥40 U/mL). ELISA was used for the measurement of anti-double-stranded deoxyribonucleic acid (dsDNA), anti-beta-2-glycoprotein I (B2GPI), anti-cardiolipin, and antiphospholipid antibodies (ORGENTEC Diagnostica GmbH, Mainz, Germany). These commercially available methods were used following the manufacturer's protocol.

2.4 Determination of COVID-19 infection

COVID-19 disease diagnosis was made by (1) positive rapid antigen test and/ or positive polymerase chain reaction testing performed with oral/nasopharyngeal swabs and/or (2) anti-N and anti-S protein antibody positivity without previous vaccination.

2.5 Statistical analysis

The statistical analysis was performed using the SPSS version 29 (IBM, Armonk, NY). The normality of distributions in the case of continuous variables was tested using the Shapiro–Wilk test. Normally distributed continuous variables were described by mean and SD values. Categorical variables were described using frequencies (case numbers) and percentages. To compare several matched groups, the Friedman test was used since data were not normally distributed (Shapiro–Wilk test: p > .05). Two related scale variables were compared, either by the paired t-test, or the Wilcoxon test, depending on the distribution. The contingency tables presenting the connection of two categorical variables were analyzed by Fisher's exact test. To identify the significant independent predictors of the relapse of myositis, forward stepwise (likelihood ratio) logistic regression analysis was applied, using the following candidate variables: age, duration of the disease, gender, disease subtype and state (PM/DM, activity at the time of vaccination), autoantibody profile, comorbidities (chronic obstructive pulmonary disease, ILD, diabetes mellitus, hypertension, ischemic heart disease, asthma), medications (glucocorticoids, methotrexate, leflunomide, cyclosporine-A (CSA), azathioprine, mycophenolate mofetil (MMF), intravenous immunoglobulin, hydroxychloroquine, cyclophosphamide, rituximab), and vaccination features (number of vaccination, manufacturers). We also calculated the odds ratios (ORs) for the predictors found. A p value less than .05 was regarded as statistically significant.

3.1 Demographic data, vaccinations, and relapse incidence

One hundred and eighty-one patients were contacted, 176 of whom finally participated in the study. The epidemiologic data of the interviewed study population are summarized in Table 1. By May 1, 2022, 82.9% (146/176) of the patients had received an anti-COVID-19 vaccination, 67% (99/146) of whom were given the Pfizer-Biontech type (BNT162b2). One hundred and thirty nine patients received both the first and the second dose, while seven patients received only one vaccination. No major immediate adverse events, such as hospitalization or anaphylaxis were reported. 11.6% (17/146) of patients who were vaccinated experienced a relapse (based on the IMACS criteria) within 3 mo of the vaccination. The general characteristics of the unvaccinated population (n = 30) showed distinct similarities and differences in comparison with the vaccinated group (Table 1). We found that the relapse rate of the unvaccinated patients was 3.3% during the follow-up period, which was not different in comparison with the relapse rate of the vaccinated group (Fisher: p = .32; OR, 3.82; confidence interval [CI], 0.49–29.9). However, due to small sample size a significant effect in either direction cannot be confidently excluded.

3.2 Features and outcomes of post-vaccination disease relapses

The major characteristics of the relapses are shown in Table 2. No patient died during their relapse, but two patients needed hospitalization, with one patient requiring intensive care because of breathing insufficiency. The average time from last vaccination to the patient reporting a worsening of symptoms was 25 days. Thirteen patients (8.9%) experienced a disease worsening within 1 mo of the last vaccination. The severity of the relapses was minimal in three (17.6%), moderate in six (35.3%), and major in eight (47.1%) patients. The CK levels were significantly different during follow-up (p = .022). In 3/17 cases, new auto-antibodies were detected after vaccination. One patient had high titer anti-aminoacyl-transfer-ribonucleic-acid synthetase antibody (anti-EJ) positivity, and two patients had an occurrence of anti-SSA antibody as a second autoantibody in addition to anti-Jo1 and anti-TIF1 gamma respectively.

After the relapses, all of the patients required different individual therapeutic interventions, such as corticosteroids, an increase in dose of a disease modifying antirheumatic drug (DMARD), administration of a new DMARD, intravenous immunoglobulin (IVIG), treatment with a biological agent or a combination of the above. The details of each relapses (symptoms, treatment schedule, scores, etc.) are presented in Table S1. Three months after the relapses, 70.6% of the patients achieved at least a minimal improvement of disease activity based on the TIS calculation (seven minor, five moderate, and zero major improvements), whereas 29.4% did not respond to the treatment. Six months after the relapses improvement was detected in 88.2% of patients (three minimal, eight moderate, and four major improvements), while in two patients disease activity did not decrease despite the treatment. One patient who did not achieve at least a minimal improvement after the treatment escalation had a mild relapse, while another patient did not respond to a change in DMARD therapy.

3.3 Breakthrough COVID infections

In the current study, we could detect a confirmed COVID-19 breakthrough infection after vaccination in 18 patients (12.3%), 17 of whom (94.4%) had a mild course of infection without hospitalization. The mean duration between vaccination and infection was 22 wk (range 2–38 wk). One of our vaccinated patients with cancer associated myositis (colorectal carcinoma with multiple liver and pulmonary metastases) without post vaccination seroconversion (undetectable antibodies against spike protein) died because of tumor progression and terminal COVID-19 infection. The COVID infection rate in the follow-up period was 20% (6/30) of the unvaccinated group. All fatal COVID-19 infections in the whole study group (n = 3) occurred only in patients whose serum antibodies against spike protein were undetectable, regardless of whether or not they had been vaccinated.

3.4 Relevant factors that could contribute to post vaccination IIM relapses

When comparing each variable individually between the patients with and without relapse, several parameters were significantly different in the two populations: vaccination status, active disease at vaccination, the number of vaccinations, double vaccination (first and second dose), BNT162b2 vaccination, the presence of any myositis specific antibodies, ILD, and treatment with methylprednisolone, MMF, and CSA. However, forward stepwise logistic regression analysis revealed that only two factors were significantly associated independently with the occurrence of relapse: the active state of myositis at time of injection and the use of the BNT162b2 vaccine (Table 3). When restricting the logistic regression analysis to those patients who were not infected with COVID (eliminating the confounding effect of COVID-19 infection), in addition to active myositis at vaccination, a diagnosis of DM was associated with relapse in logistic regression analysis. The wide CI for the OR of relapse for DM showed that our limited data set cannot confirm or exclude an effect of vaccination on DM relapse. By ordinal regression analysis, we could not find any factor associated significantly with the severity of relapse.