PROCAM based myocardial infarction risk in relation to global vascular disease risk: observations from the ARCO cohort study

Abstract

BACKGROUND: In Switzerland, risk for acute myocardial infarction (AMI) has been considered as equivalent to risk for atherosclerotic cardiovascular disease (ASCVD). This may lead to an underestimation of ASCVD risk and prevent adequate preventive measures. METHODS: We calculated correction factors for AMI risk to obtain ASCVD risk, tested predicting abilities of PROCAM/AGLA, SCORE, HerzCheck® and carotid plaque imaging (TPA) for ASCVD events in this cohort study and calculated survival curves, calibration and discrimination for ASCVD outcomes derived from PROCAM/AGLA, SCORE and TPA. RESULTS: In 2842 subjects (age 50 ± 8, 38% women), 154 (5.4%) cardiovascular events occurred (ASCVD: 41 myocardial infarctions, 16 strokes or TIAs, 21 CABG, 41 PTCA, 35 coronary artery disease [CAD]defined by invasive angiography) during a mean follow-up time of 5.9 (1–12) years. AGLA-AMI risk was well calibrated for AMI (15% underreported risk for the risk of AMI), but was poorly calibrated for ASCVD (stroke, CABG, PTCA or CAD, which contributed to the secondary outcome variables) with underreported risk resulting in a correction factor of 3.45. Discrimination was comparable for all risk calculators, but TPA outperformed risk calculators for survival using Cox proportional survival functions. Net reclassification improvement for PROCAM and SCORE using TPA tertiles groups increased significantly between 30% to 48%. CONCLUSIONS: PROCAM-derived risk calculators are well calibrated for the risk of AMI. PROCAM-AMI should be multiplied by a factor of 4 to obtain ASCVD. PROCAM-AMI does not represent global cardiovascular risk. Corresponding adjustments in the AGLA communication of risk appear necessary.