Background Colour Doppler sonography (CDS) is an emerging diagnostic tool for giant cell arteritis (GCA), however, its use in routine practice is still not widespread.

Objectives To develop a protocol using CDS and explore its value in the routine care of patients with GCA.

Methods We developed a structured scanning protocol for CDS of temporal and axillary arteries (total of 8 anatomical sites scanned per patient) based on previously published methods. We tested the protocol on consecutive patients referred to a single rheumatology centre, with suspected or established GCA, between July 2014 and September 2016. We defined a positive scan by the presence of halo in at least one branch of a temporal artery (TA) or one axillary artery (AX). We report data from the first 293 consecutively scanned cases.

Results We assessed 293 patients (mean age 72±10, female/male 196/97), of whom 118 had clinically confirmed GCA. Amongst new referrals with confirmed diagnosis of GCA, 44% had a positive scan; two patients with a positive scan did not have GCA. 82% of new referrals patients showed exclusive TA involvement, 25% concomitant AX, and 4% exclusive AX involvement. High-dose glucocorticoid therapy had already been started in 78% of these patients for an average of 17±33 days. Amongst this group, the sensitivity of CDS was 46% (95% IC 37%>55%), specificity 98% (95% CI 93%>99.8%), positive predictive value 96% (88%>99.6%), and negative predictive value 60% (95% IC 52%>68%). During the period of observation, the rate of temporal artery biopsies (TAB) decreased significantly from 42% to 25% (p=0.002). During follow up, CDS was positive in 21% of 89 routine scans in asymptomatic individuals, compared to 58% in patients with confirmed clinical flares (45% of whom had negative inflammatory markers). Over time, the number of halos per patient reduced; only new or flaring patients showed a halo in >4 sites. Halo size at the TA did not change significantly (average thickness 0.6±0.1 mm), however, the size of AX artery halos significantly reduced from first referral (1.6±0.4 mm) to follow up (1.4±0.2, p=0.01) or during subsequent flares (1.4±0.2, p=0.02).

Conclusions We have developed and tested a standardised methodology for CDS evaluation of GCA. CDS provides a high positive predictive value for a diagnosis of GCA in unselected patients from routine clinical practice, although prior high dose glucocorticoid therapy is likely to reduce its sensitivity. CDS allows for a significant reduction of TAB. We explored the role of CDS to detect disease flares and demonstrated a significant reduction in the extent of abnormalities, and of the size of halo of the AX arteries during follow up or flares. These findings could have a significant impact on the management of patients with suspected and confirmed GCA.

Disclosure of Interest None declared