Glucocorticoid discontinuation and clinical phenotypes in patients with giant cell arteritis: A retrospective cohort study from a fast track clinic
Tanja Fromberg Gorlen, Uffe Møller Døhn, Marie Celeste Heymonet, Jane Maestri Brittain, Mikkel Østergaard, Rasmus Sejersten Ripa, Peter Riis Hansen, Søren Jacobsen, Lene TerslevBackground
Previous studies have shown that many patients with giant cell arteritis (GCA) remain on glucocorticoid (GC) therapy for several years, though mostly based on older data. Identifying predictors of successful GC tapering is increasingly relevant.
Objectives
to 1) determine the probability of discontinuing GC treatment, 2) characterise clinical phenotypes within the GCA spectrum, and 3) their predictive value for successful GC discontinuation.
Design
A retrospective cohort study of all patients diagnosed with GCA in our fast track Clinic (FTC) between September 2018 and December 2021.
Methods
Clinical, biochemical and imaging data were obtained from patient records from the time of the FTC visit until 31.12.2023. The cumulative incidence of GC discontinuation was estimated using a competing risks model. Symptom- and laboratory-based phenotypes were identified through hierarchical cluster analysis.
Results
During the study period, 172 patients were diagnosed with GCA. At baseline, 26% had a prior diagnosis of GCA or polymyalgia rheumatica (PMR). The cumulative incidence of GC discontinuation was 10% at one year, 43% at two years, 58% at three years, and 69% at four and five years after the FTC visit, respectively. Patients with prior GCA/PMR had a lower probability of discontinuing GC treatment than those with new-onset disease. Four symptom-based and three laboratory-based phenotypic clusters were identified, but none of these variables predicted successful GC discontinuation. Multivariable Fine-Gray regression confirmed that previous GCA/PMR was an independent predictor of reduced probability of GC discontinuation.
Conclusions
In this real-world cohort, many patients required GC therapy for more than two years. The identified phenotypes did not predict treatment discontinuation. A previous diagnosis of GCA/PMR was associated with a persistent disease course.