Association between antibiotic use and the onset of giant cell arteritis and polymyalgia rheumatica: A nested case–control study from E3N‐European Prospective Investigation into Cancer and Nutrition

• Published in: Journal of internal medicine
• Main Authors: Pacoureau, Lucas, Barde, François, Gelot, Amandine, Elbaz, Alexis, Fournier, Agnès, Nguyen, Yann, Seror, Raphaèle
• Format: Journal Article
• Language: English
• Published: England 02.09.2025
• Subjects: antibiotic; E3N; giant cell arteritis; disease risk factor; infections; nested case–control study; trigger; polymyalgia rheumatica
• ISSN: 0954-6820; 1365-2796; 1365-2796
• DOI: 10.1111/joim.70000

To assess the association between infections, assessed by antibiotic reimbursement, and the occurrence of giant cell arteritis (GCA) and/or polymyalgia rheumatica (PMR). We conducted a nested case-control study from the French cohort E3N-European Prospective Investigation into Cancer and Nutrition, which has followed 98,995 women since 1990. Cases, defined as patients who developed GCA and/or PMR during follow-up, were matched with 20 controls on age and vital status. Infections prior to index date, defined by ≥1 antibiotic reimbursement on the medication claims reimbursement database, were compared between groups using conditional logistic regression models, adjusted for potential confounders. Different time periods before the index date and different antibiotic classes were compared. A total of 428 GCA/PMR cases (113 GCA, 232 PMR, 83 undefined) were compared to 8560 matched controls. Compared to controls, GCA/PMR cases had higher odds to have any infection in the [0-24] months prior to index date (aOR [95% CI] 1.22 [1.00-1.51]). Considering the 6-month periods prior to index date, the association was stronger when close to index date (1.18 [0.94-1.47]; 0.95 [0.75-1.19] for [0-6] and [18-24] months, respectively). This association was only found among GCA cases (1.63 [1.08-2.48] for [0-24] months), but not among PMR cases. Quinolone reimbursements were the most associated with subsequent GCA (2.07 [1.23-3.49] for [0-12] months). Compared to controls, GCA patients were at higher risk of having used antibiotics in the 24 months prior to the diagnosis. Infections or a disbalanced microbiome could act as a trigger of the disease, although a reverse causation bias cannot be excluded.