Optimization and Validation of High, Intermediate and Low Risk Cut-off Points in Risk Functions for Primary Cardiovascular Prevention. the ORESTES Project

Medical societies recommend strategies based on the identification of high-risk individuals using mathematical functions that provide risk scores for ischemic heart disease (IHD) or cardiovascular disease (CVD) at 10 years or more. Each function suggests its cut-off points (generally multiples of 5) that classify patients into low, intermediate and high/very high-risk levels. The proportion of general population classified in the low-, intermediate- and high-risk groups may vary significantly depending on the risk function used. Therefore, we postulate that cut-off points can be setup in such a way that the distribution of the population (percentile distribution) among the risk groups is the same regardless of the risk function. To determine the risk cut-off points that produce the same population distribution in the low (65%), intermediate (20%) and high-risk (15%) categories, for the most commonly IHD and CVD scores: Framingham IQ, REGICOR-IHD, FRESCO-IHD, SCORE2, FRESCO-CVD, CUORE and AHA. And to compare and evaluate their performance according to the proposed percentile distribution cut-off points. We used a population-based cohort of 816,834 residents aged 45 to 74 years in Catalonia, Spain, with a median of a following-up greater than 10 years including fatal and non-fatal cardiovascular events and all-cause fatal events. Individuals baseline characteristics, i.e. age, sex, cholesterol, diabetes, smoking status, blood pressure and lipid-lowering and hypertension treatments, were obtained from the Primary Care Center Registries. Fatal and non-fatal events were obtained from Mortality and Hospital discharge records, respectively. The number of individuals with complete data was 437,734. The cut-off points were calculated using the percentile distribution. Performance and concordance of the risk functions are assessed by the categorical Net Reclassification Index (cNRI) and the Number Needed to Treat (NNT), and by Kappa index, respectively. Table 1 shows the cut-off points in the risk (%) of IHD and CVD necessary to obtain an equivalent population distribution of 65% at low risk, 20% at intermediate and 15% at high risk with the considered functions (percentile cut off-points), and the performance measurements (cNRI and NNT). Table 2 shows the pairwise Kappa index in assigning individuals in one of the three risk groups defined by the percentile cut-off points. The ORESTES project successfully establish cut-off points in the risk of IHD and CVD calculated using 7 risk functions commonly used in Europe, which define equivalent population groups for low, intermediate and high risk.