Outcomes of COVID-19 in patients with primary systemic vasculitis or polymyalgia rheumatica from the COVID-19 Global Rheumatology Alliance physician registry: a retrospective cohort study

In this retrospective cohort study, we sourced data from the COVID-19 Global Rheumatology Alliance physician registry and the European Alliance of Associations for Rheumatology (EULAR) COVID-19 registry. These registries contain provider-reported cases of COVID-19 among patients with rheumatic diseases.

^, 

^, 

^, 

^, 

Cases are voluntarily entered by rheumatologists or other health-care providers. Data are entered directly into the global or European data entry systems, or transferred from national registries (France, Germany, Italy, Portugal, or Sweden). Patients aged 18 years and older diagnosed with COVID-19 (confirmed or presumptive) between March 12, 2020, and April 12, 2021, who had a history of primary systemic vasculitis or polymyalgia rheumatica were included. Primary systemic vasculitis included antineutrophil cytoplasmic antibody (ANCA)-associated vasculitis (granulomatosis with polyangiitis, microscopic polyangiitis, or eosinophilic granulomatosis with polyangiitis), giant cell arteritis, Behçet’s syndrome, and other vasculitides including Kawasaki disease. A text entry option was available when inputting data to the registry to provide a specific diagnosis or another diagnosis, if not listed. Data quality was assessed by the University of California (San Francisco, CA, USA) and the University of Manchester (Manchester, UK), both of which confirmed that there were no duplicates in the data entries. Given the nature of the data collected, the UK Health Research Authority and the University of California San Francisco institutional review board considered this study exempt from the need to obtain patient consent. Both institutions provided ethics approval for this study.

Data from the COVID-19 Global Rheumatology Alliance and EULAR COVID-19 registries were collected for analysis on April 15, 2020, by the GRA data analytic center at the University of California San Francisco. All patients with primary systemic vasculitis or polymyalgia rheumatica were included in the main analysis. Given disease-specific differences in treatments and risk factors for COVID-19 outcomes, subgroup analyses were done for the following specific diagnoses: giant cell arteritis, ANCA-associated vasculitis, polymyalgia rheumatica, Behçet’s syndrome, and other vasculitis.

Immunosuppressive therapies for primary systemic vasculitis at the time of COVID-19 infection were included in the analyses and categorised into groups. DMARDs were categorised as conventional synthetic DMARDs (including antimalarials, apremilast, azathioprine or 6-mercaptopurine, colchicine, cyclosporine, cyclophosphamide, leflunomide, methotrexate, mycophenolate mofetil or mycophenolic acid, sulfasalazine, and tacrolimus) and biological and targeted synthetic DMARDs (including abatacept, rituximab, anakinra, canakinumab, tocilizumab, sarilumab, infliximab, etanercept, adalimumab, golimumab, and certolizumab pegol). Rituximab, cyclophosphamide, and glucocorticoids were also analysed separately; glucocorticoids were categorised by the prednisolone-equivalent dose (0 mg/day, 1–5 mg/day, 6–9 mg/day, or ≥10 mg/day).

The primary outcome was COVID-19 outcome, assessed by use of an ordinal COVID-19 severity scale, which was defined as: (1) no hospitalisation (ie, admission to hospital); (2) hospitalisation with no supplemental oxygen; (3) hospitalisation with any supplemental oxygen or mechanical ventilation; and (4) death.

Relevant covariates included age (analysed as a continuous variable and by decade), sex (female or male), race or ethnicity (White, Black, Latin American, or other), time period (on or before June 15, 2020; June 16 to Sept 30, 2020; or Oct 1, 2020, to April 12, 2021),

comorbidities (hypertension, cardiovascular disease, diabetes, chronic kidney disease, lung disease, interstitial lung disease, or cancer), number of comorbidities (analysed as a continuous variable), body-mass index (BMI; obese [BMI ≥30 kg/m²] or non-obese [BMI <30 kg/mg²]), smoking status (ever or never smoker), disease activity, as per the physician’s global assessment (remission, low, moderate, or high or severe), and region (Europe, North America, South America, or other). Other regions included Asia, Eastern Mediterranean, South-East Asia, and Western Pacific region.

Categorical variables are reported as numbers and percentages, and continuous variables are reported as means (SD) or medians (IQR). Data were analysed by ordinal logistic regression, and associations were estimated with odds ratios (ORs) and their associated 95% CIs. Only patients with complete outcome data were included in the models. Missing data for other variables were assumed to be missing at random. Multiple imputation was performed for all models to obtain pooled estimates for disease activity, smoking, and glucocorticoid use. An overall model included sex, age, glucocorticoid use as a categorical variable (ie, prednisolone-equivalent dose categories), medication category (no DMARDs, conventional synthetic DMARDs only, biological or targeted synthetic DMARDs only, combined biological or targeted synthetic plus conventional synthetic DMARDs, rituximab only, or cyclophosphamide only), time period, number of comorbidities, smoking status, obesity (ie, a BMI of ≥30 kg/m²), disease activity, and region. Individual ordinal regression models, which included the same covariates but with different medication categories (ie, no DMARDs, methotrexate, leflunomide, IL-6 inhibitor, azathioprine, rituximab, or cyclophosphamide), were also constructed for giant cell arteritis, ANCA-associated vasculitis, and polymyalgia rheumatica. In all models, age was treated as a continuous variable by decade, and a nominal test was used to confirm that the parallel regression assumption was met. An interaction term between prednisolone usage (binary) and disease activity was included as an exploratory analysis in the overall population.

^, 

We also did a sensitivity analysis including independent comorbidities (hypertension, cardiovascular disease, diabetes, chronic kidney diseases, lung disease, or interstitial lung disease) in patients with ANCA-associated vasculitis. Results were considered statistically significant at a two-sided p value of less than 0·05. Analyses were done in R, version 4.0.2.

The funders of the study had no role in study design, data collection, data analysis, data interpretation, or writing of the report.