Carotid intima media thickness (cIMT) is an intermediate phenotype of early atherosclerosis that independently predicts vascular events. It is often suggested that cIMT be used as a screening tool to select subjects with an elevated event risk. Whether cIMT adds information to traditional risk models has so far received little investigation.

The 10-year follow-up of 4904 subjects from the Carotid Atherosclerosis Progression Study (CAPS) without pre-existing vascular disease included cardiovascular events and total mortality. Using Cox models and reclassification statistics, we investigated the usefulness of cIMT in individual risk prediction beyond the Framingham and the SCORE models, using risk strata of 0–5, 5–10, 10–20, and ≥20% over 10 years. Carotid intima media thickness was significantly and independently predictive for cardiovascular events. Compared with a model using the Framingham risk factors, a second model that included the common carotid-IMT led to the reclassification of 357 subjects (8.1%). In 107 subjects (30.0%), this reclassification was correct as confirmed with the actual outcome over 10 years. Net reclassification improvement was −1.41% (P = NS); integrated discrimination improvement was 0.04% (P = NS). More subjects were shifted to lower than to higher risk categories by the inclusion of cIMT. Analyses including other endpoint definitions, other carotid segments, and the SCORE risk model for baseline prediction did not result in consistently better risk prediction with cIMT.

Despite cIMT being predictive for cardiovascular endpoints, it did not consistently improve the risk classification of individuals. Carotid intima media thickness may not be useful for the risk stratification of individuals in the general population.