Cardiovascular Risk Prediction Equations Underestimate Risk in People Living with HIV: Comparison and Cut-point Redefinition for 19 Cardiovascular Risk Equations.

Abstract

Background: Rates of cardiovascular disease are higher in people living with HIV. Early detection of high-risk subjects (applying cardiovascular risk equations) would allow preventive actions. D:A:D, ASCVD, and FRS:CVD equations are the most recommended. However, controversies surround these equations and cut-points, which have the greatest capacity to discriminate high-risk subjects.

Objectives: The study aims (i) to assess the association/agreement between cardiovascular risk levels obtained with D:A:D and fifteen other cardiovascular risk equations, (ii) to detect cardiovascular risk equation's capability to detect high-risk subjects, and (iii) to specify the optimal cardiovascular risk equation´s cut points for the prediction of carotid plaque presence, as a surrogate of high cardiovascular risk.

Methods: 86 adults with HIV were submitted to the clinical, laboratory, and cardiovascular risk evaluation (including carotid ultrasound measurements). Cardiovascular risk was evaluated through multiple risk equations (e.g., D.A.D, ASCVD, and FRS equations). Association and agreement between equations (Correlation, Bland-Altman, Williams´test) and equation's capacity to detect plaque presence (ROC curves, sensitivity, specificity) were evaluated.

Results: Cardiovascular risk equations showed a significant and positive correlation with plaque presence. Higher high-cardiovascular risk detection capability was obtained for ASCVD and D:A:D. Full D:A:D5y>0.88 %, ASCVD>2.80 %, and FRS:CVD>2.77 % correspond to 80 % sensitivity.

Conclusion: All cardiovascular risk equations underestimate the true risk in HIV subjects. The cut-- points for high cardiovascular risk were found to vary greatly from recommended in clinical guidelines.