COVID-19 ADMISSIONS AND MORTALITY IN PATIENTS WITH EARLY INFLAMMATORY ARTHRITIS: RESULTS FROM A NATIONAL COHORT

ABSTRACT

Background/Aims Patients with inflammatory arthritis were identified as a potentially vulnerable group during the COVID-19 pandemic, with recommendations from the UK government to shield. We set out to describe the risks of COVID-19 according to initial treatment strategy amongst patients recruited to the National Early Inflammatory Arthritis Audit (NEIAA). Methods NEIAA is an observational cohort design. It includes adults in England with a new diagnosis of inflammatory arthritis between May 2018 and March 2021. The outcomes of interest were death due to COVID-19 (COVID-19 stated on a death certificate) and hospitalisation due to COVID-19 (primary admission reason or nosocomial acquisition), identified using NHS Digital linkage. Cox proportional hazards models were used to calculate hazard ratios, with adjustment for patient factors (age, gender, smoking status, comorbidity) and disease factors (seropositivity, disease severity (DAS28), patient-reported disability (HAQ) and functional impact (MSK-HQ)) recorded at baseline. Individuals were considered at risk from February 2020 or date of diagnosis (whichever was later) and censored at a COVID-19 event, May 2021 or death (whichever was sooner). Results 14,127 patients were included. Mean age was 57 (+/-16);62% were female. Smoking status 19% current;29% ex-smokers. Comorbidities 19% hypertension;9% diabetes;and 9% lung disease. Overall, 20% had two or more comorbidities. Rheumatoid Factor or CCP antibodies were positive in 56%. At presentation, mean scores were 4.6 (+/-1.5) for DAS28, 1.1 (+/-0.7) for HAQ and 25 (+/-11) for MSK-HQ. Initial DMARD therapy was known for 13,682/14,127 patients;methotrexate was most common (54%), then hydroxychloroquine (23%) and sulfasalazine (11%). There were 143 COVID-19 hospital admissions and 47 deaths, corresponding to incidence rates per 100 person-years for hospitalisation 0.94 (95% CI 0.79-1.10) and death 0.31 (95% CI 0.23-0.41). Increasing age, male gender, diabetes, hypertension, lung disease and smoking status all predicted COVID-19 events. Higher baseline DAS28 predicted COVID-19 admission (HR 1.24 (95% CI 1.10-1.39)) and mortality (HR 1.33 (95% CI 1.09-1.63)). Higher HAQ predicted both COVID-19 admission and death. Seropositivity was not a significant predictor of any COVID- 19 event, nor was MSK-HQ. Unadjusted, corticosteroids associated with COVID-19 death (HR 2.29 (95% CI 1.02-5.13)), and sulfasalazine monotherapy associated with COVID-19 admission (HR 1.93 (95% CI 1.04-3.56)). In adjusted models, associations for corticosteroids and sulfasalazine were no longer significant. Only age, smoking status, and comorbidities independently predicted COVID-19 events. Conclusion The burden of COVID-19 amongst early arthritis patients was substantial during the pandemic. Patient characteristics and rheumatoid disease severity at diagnosis appear to be the more important predictors of COVID-19 events than initial treatment strategy. An important limitation is that we have not looked at treatment changes over time, and must acknowledge that many patients, especially those recruited in 2019, may have changed therapy prior to the pandemic.