[Validation of 3 equations of coronary risk in diabetic population of a primary care center].

BACKGROUND AND OBJECTIVE

We intended to validate to 10 years 3 equations of coronary risk that use the function of original Framingham (Framingham-Anderson and Framingham-Wilson) and calibrated for Spanish population (Framingham-REGICOR) in diabetic patients type 2.

PATIENTS AND METHOD

We included a total of 190 diabetic patients from a primary care center, without coronary heart disease, in whom the coronary risk could be calculated before 1 January 1995. All were followed during 10 years.

RESULTS

The highest score of cardiovascular risk was obtained in the Framingham-Wilson function chart (25.4%) and the lowest in the Framingham-REGICOR (10.8%). The real incidence of coronary events was 14.7% (p < 0.001). Statistically significant differences between patients with or without coronary events were only observed in the Framingham-REGICOR equation (13.3% vs 10.3%; p = 0.046). Framingham-Anderson and Framingham-Wilson functions charts more than doubled the risk of coronary disease in men (p < 0.001). The agreement degree between the 3 functions was acceptable except for the calculation of coronary risk in men between Framingham-REGICOR and Framingham-Wilson equations (kappa index = 0.3). Framingham-Anderson and Framingham-Wilson functions charts showed similar profiles for the patients with high coronary risk. Sensitivity was superior in Framingham-Anderson and Framingham-Wilson functions (67.8%) and specificity was optimal in Framingham-REGICOR equation (51.2%). The positives predictive values were low and the negatives predictive values were high.

CONCLUSIONS

The equations of Framingham-Wilson and Framingham-Anderson overestimate the coronary risk in diabetics, whereas the Framingham-REGICOR functions underestimates it. The utility of these 3 methods is reduced in the diabetic population.