Diagnostic validity of ultrasound including extra-cranial arteries in giant cell arteritis.

OBJECTIVES

Color Doppler ultrasound (CDUS) of the temporal arteries (TA) is becoming the first test to be performed for suspected giant cell arteritis (GCA). Our aim was to assess the added value of including CDUS of large vessels (LV) in the diagnosis of GCA.

METHODS

We performed an observational and retrospective study of consecutive patients with suspected GCA. Baseline CDUS of the TA and LV (axillary, subclavian, and carotid) were conducted. We defined the CDUS finding as positive if the halo sign was present.

RESULTS

Of 198 patients with suspected GCA, 87 were eventually diagnosed with GCA: 45 (51.7%) had a cranial pattern exclusively, 31 (35.6%) had both a cranial and an LV pattern, and 11 (12.6%) had an isolated LV pattern. CDUS of the TA had a sensitivity of 83.9%, specificity of 97.3%, and positive and negative predictive values (PPV, NPV) of 96.1% and 88.5%, respectively. When LV was added, sensitivity increased to 96.6% and NPV to 98.2%. Specificity was 97.3% and PPV was 96.6%. As for LVs, the axillary, subclavian, and carotid arteries were involved in 87.8%, 77.4%, and 34.4%, respectively. Isolated axillary examination resulted in a loss of 12.2% of patients with LV involvement; however, inclusion of the axillary and subclavian arteries retained 100% of patients with LV involvement.

CONCLUSIONS

Detection of GCA by ultrasound should routinely include examinations of the TA and LV (at least the axillary and subclavian arteries) to improve diagnostic accuracy. More than 12% of patients in our cohort had isolated LV involvement. Key Points • Extracranial involvement in GCA is very common: half of patients have extracranial vasculitis and more than 12% isolated LV involvement that can be demonstrated with CDUS. • Adding a CDUS examination of LV to TA increased sensitivity (from 83.9 to 96.6%) and the negative predictive value (from 88.5 to 98.2%) for diagnosis of GCA. • In our cohort, if we only examined the axillary arteries, 12.2% of the CGA with LV involvement would not have been diagnosed. • We propose a CDUS protocol that includes examination of the TA and LV (at least the axillary and subclavian arteries) routinely in cases of suspected GCA.