• José Miguel Baena Díez, José Luis del Val Garcia, Luis Héctor Salas Gaetgens, Ricard Sánchez Pérez, Eva Altes Vaques, Bibiana Deixens Martínez, Marta Amatller Corominas, and Dynaida Katia Núñez Casillas.
• Area Básica de Salud, Dr Carles Ribas-La Marina, Barcelona. jbaenad@meditex.es
• Rev Esp Salud Publica. 2005 Jul 1; 79 (4): 453-64.

BackgroundNo studies have been published to date comparing SCORE, REGICOR and Framingham models. This study is aimed at analyzing how the REGICOR and SCORE functions classify cardiovascular risk, their correlation and concordance with Framingham (1998) and whether any differences exist among them with regard to the cardiovascular risk factors in high-risk groups.MethodsDescriptive cross-sectional study conducted in primary care. A total of 851 individuals within the 35-74 age range, free of cardiovascular diseases and selected by simple random sampling were included. A study was made of the percentage of high-risk patients with Framingham (> or = 20% ten-year risk), SCORE (> or = 5% ten-year risk) and REGICOR with cutoff points > or = 20%, > or = 15%, > or = 10% and > or = 5% at 10 years, given that with REGICOR > or = 20% there are hardly any high-risk cases. A comparison was drawn between the correlation (Pearson's r) and concordance (Kappa index) of the REGICOR and SCORE high-risk individuals as compared to Framingham.ResultsThe high-risk percentages respectively found were: 23.3% with Framingham; 15.2%; with SCORE; and 1.4%, 5.8%, 17.6% and 57.0% with REGICOR with the cutoff points described. REGICOR has a 0.99 correlation, SCORE a 0.78 correlation. REGICOR > or = 10% showed a better concordance (Kappa 0.83) than SCORE (Kappa 0.61). On comparing the cardiovascular risk factors of the high-risk cases (> or = 20% Framingham, > or = 5% SCORE and > or = 10% REGICOR), SCORE showed higher prevalence of diabetes and a lower prevalence of hypercholesterolemia (p<0.05).ConclusionsREGICOR showed a good correlation with Framingham. With the > or = 10% cutoff point, it classifies a number of individuals as high-risk similar to SCORE and fewer than Framingham. The SCORE model would treat a number of patients similar to the REGICOR > or =10% model with hypolipemiant drugs, however showing lesser evidence of effectiveness of the treatment.

13 keywords...

hide…