Agreement among available cardiovascular disease (CVD) prediction models in people living with HIV (PLWH) remains unclear. This study evaluated the agreement of commonly used 10-year CVD prediction models in older Chinese PLWH and examined model fairness across socioeconomically disadvantaged populations. In this cross-sectional study, 10-year CVD risk was calculated using the Data collection on Adverse Effects of Anti-HIV Drugs Study risk equation (D:A:D; 2010), the reduced and full version of D:A:D (2016), the Framingham Risk Score (FRS), the Systematic COronary Risk Evaluation (SCORE) for the low- and high-risk region versions, the pooled cohort equations of the American Heart Society/American score (PCE), and the Predicting Risk of Cardiovascular Disease EVENTs (PREVENT). Agreement between models was assessed using weighted Kappa (κ) statistics, with subgroup fairness analyses by sex, age, and employment status. Among 473 individuals with a mean age of 62.03 (SD = 7.64) years, the observed agreement between D:A:D (2010) and the other models ranged from 34% for SCORE (low) to 77% for PCE. The difference in 10-year CVD risk between the <60 and ≥60 years groups assessed with D:A:D (2010) was greater than the difference with other models. Agreement between D:A:D and the FRS, SCORE, PCE, and PREVENT was poor. Future HIV-specific CVD risk models should incorporate HIV-related factors and social determinants of health and be developed and validated for Chinese populations.
Yu et al. (Thu,) studied this question.