Cranial ultrasound for the diagnosis of giant cell arteritis. A retrospective cohort study.

BACKGROUND: Establishing a diagnosis of giant cell arteritis, or indeed ruling it out, may be difficult. We describe an evaluation of temporal artery colour duplex ultrasound as first line investigation in patients with suspected giant cell arteritis. METHODS: A retrospective cohort study of all patients undergoing colour duplex ultrasound for suspected giant cell arteritis between January 2005 and January 2014 was undertaken at a teaching hospital. A minimum clinical follow-up of three months was required. Patients were classified on the basis of ultrasound reports, using described features such as a halo sign or arterial wall thickening and clinical diagnosis of giant cell arteritis after at least 3 months follow-up, determined by the treating physician. The relationship of colour duplex ultrasound to a final clinical diagnosis of giant cell arteritis was analysed. RESULTS: A total of 87 patients underwent colour duplex ultrasound: 36 (41%) had clinically confirmed giant cell arteritis at 3-month follow-up. The positive predictive value of colour duplex ultrasound for a clinical diagnosis at 3 months was 97% (95% confidence interval (CI) 93 to 99%) and negative predictive value 88% (95% CI 76 to 95%). Sensitivity was 81% (95% CI 64 to 92%) and specificity 98% (95% CI 90 to 100%). CONCLUSIONS: A high positive and negative predictive value of arteritis on colour duplex ultrasound indicates that temporal artery biopsy may be unnecessary in suspected giant cell arteritis, particularly where clinical suspicion of giant cell arteritis is high or low.