Background: Carotid intima-media thickness (CIMT) serves as an indicator of atherosclerosis and cardiovascular risk in observational and intervention studies. Off-line measurements from stored B-mode images using manual tracing or automated-edge-detection programs are the most applied methods. Direct measurements by automated radio-frequency (RF) approach might be an interesting alternative. We compared these methods in terms of risk-factor relations and associations with future events.

Methods: Data from participants of the SMART-study was used. Far wall common CIMT was measured with B-mode and automated RF. Detailed risk factor information was obtained. All participants were followed for occurrence of vascular events (mean follow-up 2.1 years). CIMT was related to risk factors with linear-regression models and to future events with Cox-Proportional-Hazards models.

Results: Data were available for 2146 participants. Intraclass correlation between two methods was modest (0.45). The relation between B-mode CIMT with age and systolic blood pressure was twice as strong as compared to RF CIMT. The relation of B-mode CIMT with events was stronger than for RF CIMT: vascular death (1.27 vs. 1.00) and ischemic stroke (1.45 vs. 1.03). In subjects with B-mode measured CIMT < 1.00 mm, RF CIMT showed stronger relationships with vascular death (1.30 vs. 0.80), although B-mode CIMT was stronger associated with ischemic stroke (3.70 vs. 0.97).

Conclusion: Given our findings, the choice for either B-mode CIMT or RF CIMT measurements in research is partly driven by type of study-population, expected presence of local atherosclerotic abnormalities, and of the main aim of the study (risk-factors or events).