Is duplex ultrasonography useful for the diagnosis of giant-cell arteritis?

Background: Evidence of a dark halo on ultrasonography has been considered a specific sign of giant-cell arteritis and may replace temporal artery biopsy for the diagnosis of giant-cell arteritis in patients with typical clinical manifestations.

Objective: To assess the usefulness of temporal artery duplex ultrasonography and to compare this mode of ultrasonography with physical examination of temporal arteries for the diagnosis of giant-cell arteritis in patients with suspected giant-cell arteritis or polymyalgia rheumatica.

Design: Diagnostic test study.

Setting: Several divisions of Reggio Emilia Hospital, Reggio Emilia, Italy.

Patients: 86 consecutive patients with a suspected diagnosis of giant-cell arteritis or polymyalgia rheumatica identified over a 22-month period.

Measurements: The temporal arteries were examined in all 86 patients. Duplex ultrasonography of the temporal arteries was then performed by two ultrasonographers who were unaware of the clinical diagnosis. Before corticosteroid therapy was started, temporal artery biopsies were performed in all patients at the site targeted by the ultrasonographer.

Results: A hypoechoic halo around the lumen of the temporal arteries had a sensitivity of only 40% (95% CI, 16% to 68%) and a specificity of 79% (CI, 68% to 88%) for the diagnosis of biopsy-proven giant-cell arteritis. The negative likelihood ratio was 0.8 (CI, 0.5 to 1.2), and the positive likelihood ratio was 1.9 (CI, 0.9 to 4.1). When the thickness of the halo was at least 1 mm, specificity increased to 93% (CI, 84% to 98%) and the positive likelihood ratio increased to 5.7 (CI, 2.0 to 16.2); however, sensitivity remained low at 40% (CI, 16% to 68%). On physical examination, temporal artery abnormalities had a higher sensitivity of 67% (CI, 38% to 88%), a higher specificity of 99% (CI, 92% to 100%), and a higher positive likelihood ratio of 47.3 (CI, 6.5 to 342.4) than did ultrasonographic findings. None of the patients with giant-cell arteritis had a normal temporal artery inspection and a hypoechoic halo on ultrasonography.

Conclusion: Evidence on ultrasonography of a halo around temporal arteries, either any halo or a halo 1 mm or greater in thickness, only modestly increased the probability of biopsy-proven giant-cell arteritis but did not improve the diagnostic accuracy of a careful physical examination.