Carotid Intima-Media Thickness Manual Measurements: Intraoperator and Interoperator Agreements Under A Strict Protocol in a Large Sample.

PURPOSE: The aim of this study was to assess the intraoperator and interoperator agreement for manual measurements of intima-media thickness (IMT) performed under a strict carotid ultrasound technical protocol. METHODS: Two blinded experienced operators independently performed an ultrasound examination at the distal common carotid of 242 subjects in the same patient's position, diastolic phase, probe type, zooming, and depth. Thirty-six subjects were reevaluated in another time point. Three different-angle manual measurements (IMTindiv) were obtained. Interoperator agreements for each IMTindiv, and their mean (IMTmean) and maximum (IMTmax) values, were assessed with the intraclass correlation coefficient and Bland-Altman analysis. Intraoperator agreement was tested taking advantage of the second ultrasound round in 36 subjects. RESULTS: IMTmean agreements (intraoperator, 0.665-0.913; interoperator, 0.856-0.897) were higher than IMTmax (intraoperator, 0.435-0.793; interoperator, 0.631-0.718) and any IMTindiv (intraoperator, 0.355-0.676; interoperator, 0.590-0.717). Despite the small systematic error for IMTmean (intraoperator, ≤0.03; interoperator, ≤0.02 mm), at best of times, the sampling error size reached at least 0.28 and 0.25 mm for intraoperator and interoperator agreements, respectively, and was never less than 0.13 mm. CONCLUSIONS: Although IMTmean agreement is excellent under a strict protocol, limits of agreement might be too wide to consider carotid ultrasound a robust cardiovascular risk biomarker.