Background Temporal artery ultrasound (TAUS) is increasingly used in the diagnosis of giant cell arteritis (GCA). Vessel wall oedema (Halo) is suggestive of GCA. However, its prognostic importance and association with temporal artery biopsy (TAB) remain unclear. We report a semi-quantitative composite score for vascular oedema (Halo Score), its ability to predict positive histology on TAB, and the association between ultrasound characteristics and clinical and laboratory features in GCA.
Objectives To investigate whether the extent & severity of temporal and axillary artery oedema (Halo Score, HS) or the number of affected arterial branches is associated with clinical and laboratory characteristics and TAB findings in newly diagnosed GCA.
Methods We conducted a single centre retrospective data analysis from 91 consecutive patients (mean age 72.6 years, 72.5% female) with GCA according to ACR criteria, who underwent ultrasound of bilateral temporal and axillary arteries by a single blinded sonographer. The branches of the vessels were systematically assessed for the Halo sign. This was then semi-quantitatively scored at each site as no halo (score 0), the smallest third (1st tertile) of halo at each branch as mild (score 1) halos, 2nd tertile as moderate (score 2) and the largest third of halo (3rd tertile) as extensive (score 3), as the extent of the halo sign is linked to the diameter of the arterial branch. Total Halo Score (HS) was constructed as the sum of grades from all sites. TAB results were available from 79 (86.8%) patients. For statistical analysis, Mann-Whitney U, Spearman's rank correlation, Chi-square tests and ROC curve analysis were applied.
Results Halo was detected in 57 (62.6%) patients. ESR (37 vs 20mm/h, p=0.018) and CRP (27 vs 5mg/l, p=0.039) were higher in patients with positive halo. TAUS positive patients were more commonly male (35.1% vs 14.7%, p=0.035), and more often had scalp tenderness (75.4% vs 41.2%, p=0.001), constitutional upset (68.4% vs 44.1%, p=0.022) and positive TAB (40.4% vs 7.4%, p=0.002). They were less likely to have visual symptoms (15.8% vs 38.2%, p=0.016). No difference was found in headache, temporal swelling or pulsation, jaw or limb claudication, anaemia and polymyalgic symptoms.
In TAUS positive patients, the median number of affected branches was 2 (range 1-6) and median HS 4 (range 1-18). Association with patient characteristics are given in the table.
There was good agreement between HS and number of affected branches. HS also correlated with age (corrcoeff 0.29, p=0.028), ESR (0.27, p=0.04) and CRP levels (0.36, p=0.005).
The relationship between positive TAB and HS was assessed using ROC curve analysis yielding AUC 0.81 (95% CI 0.68-0.93). A Halo Score of 4 predicts positive TAB with high sensitivity (90.5% sensitivity, 71.0% specificity) and HS 7 gives high specificity (sensitivity 42.9%, specificity 90.3%).
Conclusions Our findings support vascular ultrasound as a useful non-invasive investigation in GCA. Halo findings were associated with demographic, clinical and laboratory features in GCA. Halo score was able to predict positive TAB findings. A high HS may replace the need for diagnostic TAB in some cases.
Disclosure of Interest None declared