Cardiovascular Risk Prediction Models in People Living with HIV in Colombia

Abstract

BACKGROUND: People living with HIV are at increased risk of cardiovascular disease. Cardiovascular risk (CVR) prediction scores are powerful tools for individualized assessment that inform decision-making about follow-up frequency, hypolipemiant treatment intensification, and choice antiretroviral therapy. OBJECTIVES: The objectives of the study were to evaluate the performance of multiple cardiovascular assessment scores in predicting major adverse cardiovascular events (MACE) at 5 and 10 years. Framingham (2004, 2008, and Colombia-adjusted), SCORE, PROCAM, ASCVD, and D:A:D scores were included in the analysis. METHODS: Data were obtained from a medical registry of adults living with HIV attended by a teaching hospital in Colombia. All patients with complete information necessary for risk score calculations and determination of MACE at 5 and 10 years were included in the study. Receiver operating characteristic curves (ROC) were generated using calculations with all the aforementioned models for every individual. Differences between curves were compared with De- Long's test. RESULTS: A total of 808 patients were included in the analysis. Mean age was 35 years, and 12% were female. The majority of subjects had low and very low CVR. Eight MACE occurred during follow-up. Area under ROC curves were: Framingham (0.90), Framingham ATP3 (0.92), Framingham calibrated for Colombia (0.90), SCORE (0.92), PROCAM (0.92), ASCVD (0.89), and D:A:D (0.92), with no statistically significant differences. CONCLUSIONS: The evaluated scores had an acceptable performance for HIV-infected patients in the studied cohort, especially for those in low and very low risk categories.