Predictors of giant cell arteritis in patients with polymyalgia rheumatica in southern Sweden-a retrospective study.

Objective: To identify predictors of GCA in patients with PMR. Methods: Patients with PMR were identified among participants in two population-based health surveys. Those with a registered diagnosis indicating PMR in national and regional registers (the latter including primary care) were included. Medical records from hospitals and primary care were systematically reviewed. PMR diagnoses were verified by a rheumatologist in an independent review. Potential predictors were examined using logistic regression analysis. Results: Of 1508 medical records, 1030 had sufficient data available. PMR diagnoses were verified in 626 patients (61%). GCA was diagnosed within 1 month of PMR diagnosis in 37 patients and at a later time point in 20 patients. Female patients were more likely to develop GCA at PMR diagnosis or later [odds ratio (OR) 2.38 (95% CI 1.23, 4.61)]. Higher ESR and CRP levels were also associated with GCA. A lower risk for GCA was seen in those presenting with pain/stiffness in the hip [OR 0.51 (95% CI 0.28, 0.92)]. Conclusion: In this large cohort of patients with verified PMR, GCA was diagnosed in a limited subset (9%) and was more common in females. A lower risk for GCA was seen in patients with pain/stiffness in the hip at onset of PMR, suggesting that prominent musculoskeletal symptoms and cranial arteritis represent different parts of the GCA-PMR spectrum.