A longitudinal observational study investigated whether the measurement, in clinical practice, of carotid maximum intima-media thickness (Max-IMT) could be combined with the Framingham risk score (FRS) to improve the predictability of cardiovascular events in dyslipidemic patients who are at low or intermediate risk. Max-IMT was measured by ultrasound in 1969 patients attending a lipid clinic. The "best threshold values" (BTVs) above which we considered the Max-IMT to be abnormally high were calculated for our dyslipdemic population for each 10-year age interval in men and women. Two hundred and forty-two patients (age 54+/-10 years; 43.8% women) with an FRS <20%, i.e. at low or intermediate risk, were monitored for more than 5 years. Twenty-four of these patients suffered a cardiovascular event within 5.1+/-2.3 years. Both FRS and Max-IMT proved to be independent outcome predictors (p<0.04, both), with a hazard ratio (HR) of 6.7 (95% CI 1.43, 31.04; p=0.015) in patients in whom FRS was 10-20% and Max-IMT was above the BTV (60th percentile of Max-IMT distribution for men or 80th for women). In Kaplan-Meier analysis, the Max-IMT significantly improved the predictive value of the FRS (chi(2)=8.13, p=0.04). Patients with FRS 10-20% (currently considered intermediate-risk) and also elevated Max-IMT values came into the same high-risk category as patients with FRS 20-30%. The combination of FRS with Max-IMT measurement can be used in routine clinical practice to greatly enhance the predictability of cardiovascular events in the large number of patients who fall into the intermediate-risk category, which currently does not call for aggressive preventive measures.