Scoping review of clinical, laboratory, histopathological and imaging predictors of relapse in adults with biopsy- or temporal artery ultrasound-confirmed giant cell arteritis on standard therapy with glucocorticoids alone or in combination with biologics.

OBJECTIVE: This literature review aims to summarise current evidence on clinical, laboratory, histopathological and imaging predictors of relapse in adults with biopsy- or imaging-confirmed GCA. METHODS: A comprehensive search of Ovid MEDLINE ALL and Embase databases identified 1030 articles. After screening, 58 studies met inclusion criteria. Data were extracted and thematically analysed across four domains: clinical, laboratory, imaging and histopathology characteristics. RESULTS: Relapse occurs in >40% of patients on glucocorticoid monotherapy and ≈20% on tocilizumab, most commonly within 2 years from diagnosis. PMR symptoms and headache are the most common clinical manifestations at relapse. Laboratory markers such as ESR and CRP are widely used but lack specificity. Higher inflammatory burden at diagnosis, reflected by anaemia or thrombocytosis, correlates with future relapse. Emerging biomarkers-including calprotectin, osteopontin, IL-6, angiopoietin-2 and circulating T peripheral helper cells-show promise but require validation. Temporal artery ultrasound findings such as higher baseline halo counts, delayed ultrasound remission and limited improvement in quantitative scores are the most reliable relapse predictors. Fluorodeoxyglucose PET has limited value in predicting relapse. MRI vessel wall enhancement may correlate with relapse, but evidence remains limited to small cohorts. Histological features such as greater arterial inflammation, higher giant cell density and intraluminal thrombosis are linked to a more relapsing disease course. CONCLUSION: Relapse in GCA is multifactorial and cannot be predicted by a single parameter. Integrating clinical, laboratory, imaging and histological data may improve risk stratification. Prospective studies are needed to validate predictive models and enable personalised treatment approaches.