RESUMO
OBJECTIVE: To determine the concordance and predictive capacity of various methods for calculating coronary risk in diabetic patients. DESIGN: Prospective study of cohorts, with a 10-year follow-up. SETTING: Urban health centre in a socially depressed area, with high prevalence of DM2. PARTICIPANTS: Population diagnosed with type-2 diabetes mellitus in 1991-93 (112 individuals, with an average age of 66.7+/-11.6 years, 59% of whom were male). MAIN MEASUREMENTS: At the moment of diagnosis, the parameters normally used to calculate coronary risk were determined. The tables of the Framingham scale '91 and '98, of the United Kingdom Prospective Study (UKPDS) (based on an exclusively diabetic population) and of the REgistre GIroni del COr (REGICOR-Girona Heart Register) (based on a Mediterranean population) were each used to calculate individually coronary risk at 10 years. The coronary events suffered in the 10 years following DM2 diagnosis were recorded. The Kappa indices for concordance of the tables and their predictive capacity were calculated. RESULTS: 18.2% of men and 15.2% of women suffered some coronary event. The coronary risk calculations were, for men and women, 30%-20% (Framingham '91), 36%-21% (Framingham '98), 38%-23% (UKPDS), and 15%-10% (REGICOR). CONCLUSIONS: The various methods for calculating coronary risk in diabetics suffer large variability. We should highlight their low diagnostic value in individual cases, with sensitivities of 25%-75% and low specificity (mainly in men) regarding real risk of coronary disease.