Tailoring Cardiovascular Risk Predictions for People Living with HIV: evidence from Malaysia

Background People living with HIV (PLHIV) face greater cardiovascular risk but lack targeted prevention, especially in Asia. This study compares cardiovascular risk models among PLHIV in Malaysia. Methods We included PLHIV aged ≥18 years between January 2016 and December 2017. Those with pre-existing cardiovascular disease (CVD) or non-Malaysians were excluded. Events were first cardiovascular hospitalization or death within five years. Four models were assessed: Framingham risk score (FRS), Revised Pooled Cohort Equations (RPCE), Data Collection on Adverse Effects of Anti-HIV Drugs (D:A:D) full and reduced. Discrimination was assessed with c-statistics and calibration with observed-to-expected (O/E) ratios. Results Among 6,339 PLHIV, 136 (2.1%) developed CVD. Those with CVD more often had dyslipidemia (22.8% vs. 16.8%), hypertension (23.5% vs. 7.2%), and diabetes (15.4% vs. 4.2%). All models showed similar moderate discrimination (c-statistics: 0.75-0.77, 95% CI: 0.71-0.81). D:A:D reduced calibrated better (O/E: 1.13, 95% CI: 0.94-1.32) than the full version (1.25, 95% CI: 1.04-1.46). Framingham risk score (FRS) (0.58, 95% CI: 0.49-0.68) and RPCE (0.72, 95% CI: 0.58-0.90) overestimated risk. Conclusions While all models showed similar discrimination, D:A:D reduced provided the most accurate 5-year estimates with greater practicality. This supports its use for scalable cardiovascular risk assessment in HIV care.