A reliable and safe diagnostic procedure for vertebral artery (VA) stenosis is needed, but none is generally accepted yet. In our study, we evaluated symptomatic VA stenoses using color Doppler sonography (CDS). CT angiography (CTA) has been employed as a non-invasive reference method. Next, we tested the accuracy for medium to high degree stenoses by digital subtraction angiography (DSA).

Sixty-two symptomatic patients with a proximal VA stenosis were examined prospectively with CDS and CTA. The VA diameters by both methods were correlated. The stenotic peak systolic velocity (PSV1) and its ratio to the post-stenotic segment (PSVr) were analysed using receiver operating characteristic curve (ROC). Cut-off values for PSV1 and PSVr defining moderate and severe stenosis were assessed. In stenoses≥50%, an additional search for correlation with DSA was carried out.

Mean VA diameter was 3.561mm (95% CI 3.361–3.760) by CDS and 4.180 (95% CI, 3.950–4.411) by CTA, accompanied with significant similarity in Pearson’ correlation (0.847, P<0.001). The PSV1 and PSVr appeared to be equally accurate for VA stenoses of 50% or more (PSV 1–AUC 0.814, P<0.001, cut-off velocity≥1.35m/s, PSVr–AUC 0.819, P<0.001 with a cut-off value≥2.2). Final Spearman’ correlation of CTA results vs DSA was highly significant (0.823, P<0.001).

Our results endorse the non-invasive combination of CDS with CTA in the evaluation of VA stenosis as a reliable diagnostic algorithm, tightly correlating with DSA.