Background: Colour Doppler sonography (CDS) is an established technique in the diagnosis of giant-cell (temporal) arteritis (GCA). The predictive value of its diagnostic criteria for GCA (halo sign or stenosis) is related to the pretest probability (PTP), a measure of probability of presence of a target disease before the result of a diagnostic test is known.
Patients and methods: A total of 182 (average age 69 years, 69% women) patients of the Rheumatology Center Baden-Württemberg were investigated. Based on the diagnostic criteria of the American College of Rheumatology (ACR) they were assigned to one of three groups, before a CDS was performed: group 1 (n= 139) patients with "isolated" polymyalgia rheumatica (PMR) and a low PTP for GCA; group 2 (n=19) patients with intermediate PTP and nonspecific headache and fewer than three ACR criteria for GCA); and group 3 (n=224) patients with a high PTP and new headache loclized to the temporal artery and at least three ACR criteria for GCA.
Results: The halo sign (periluminal dark halo) of more than 0.3 mm was present in 26% of group 1. 42% of those in group 2 and 83% of those in group 3. A stenosis or occlusion of the temporal artery was present in 3.5% (group 1), 5% (group 2) and 46% (group 3), respectively. 3 of 24 patients of group 3 also had a stenosis of the axillary or brachial artery. Concordance between clinical criteria and CDS (normal CDS in patients with PMR but no headache or abnormal CDS and clinically suspected BCA was found in 123 of 182 patients (67.5%). In these patients biopsy of the temporal artery ("gold standard" for the diagnosis of GCA) was not recommended. Temporal artery biopsy was, however, recommended in all patients with discordant findings (abnormal CDS with PMR but no headache or normal CDS with clinically suspected GCA, and also those with intermediary PTP (32%). A biopsy was performed in 42 of these patients after informed consent had been obtained. This demonstrated vasculitis in 11 of 25 patients with PMR (PPV in group 1: 0.44). But biopsies were negative in all four patients with clinically suspected GCA and normal CDS (NPV in group 3:1). In the intermediary group biopsy demonstrated vasculitits in 5 of 6 patients with an abnormal CDS (PPV 0.63), while 4 of 5 patients with a normal CDS had a normal biopsy (NPV 0.8).
Conclusion: Taking into account pretest probability, an RCA can be accurately diagnosed or excluded by CDS in two thirds of patients without biopsy. When performed by an experienced investigator CDS is a basic part in the diagnosis of CDA.
Copyright Georg Thieme Verlag KG Stuttgart . New York.