Performance of the pooled cohort equations and D:A:D risk scores among individuals with HIV in a global cardiovascular disease prevention trial: a cohort study leveraging data from REPRIEVE.

Abstract

BACKGROUND Risk estimation is an essential component of cardiovascular disease prevention among people with HIV. We aimed to characterise how well atherosclerotic cardiovascular disease (ASCVD) risk scores used in clinical guidelines perform among people with HIV globally. METHODS In this prospective cohort study leveraging REPRIEVE data, we included participants aged 40-75 years, with low-to-moderate traditional cardiovascular risk, not taking statin therapy. REPRIEVE participants were enrolled from sites in 12 countries across Global Burden of Disease Study (GBD) regions. We assessed the performance of the pooled cohort equations (PCE) risk score for ASCVD and the data-collection on adverse effects of anti-HIV drugs (D:A:D) risk score. We calculated C statistics, observed-to-expected (OE) event ratios, and Greenwood-Nam-D'Agostino goodness-of-fit (GND) statistics, overall and in subgroups by race, sex, and GBD regions (clustering low-income and middle-income countries and high-income countries). We did a recalibration for PCE risk score among people with HIV in high-income countries. REPRIEVE was registered with ClinicalTrials.gov, NCT02344290. FINDINGS We included 3893 participants, recruited between March 26, 2015, and July 31, 2019. The median age was 50 years (IQR 45-55), with 2684 (69%) male and 1209 (31%) female participants. 1643 (42%) were Black or African American, 1346 (35%) participants were White, 566 (15%) were Asian, and 338 (9%) were recorded as other race. Overall, discrimination of the PCE risk score was moderate (C statistic 0·72 [95% CI 0·68-0·76]) and calibration was good (OE event ratio 1·11; GND p=0·87). However, calibration suggested overprediction of risk in low-income and middle-income countries and corresponding underprediction in high-income countries. When restricted to high-income countries, we found underprediction (OE event ratio >1·0) among women (2·39) and Black or African American participants (1·64). Findings were similar for the D:A:D risk score (C statistic 0·71 [0·65-0·77]; OE event ratio 0·89; p=0·68). Improved calibration of the PCE risk score in high-income countries was achieved by multiplying the original score by 2·8 in Black or African American women, 2·6 in women who were not Black or African American, and 1·25 in Black or African American men. INTERPRETATION Among the global cohort of people with HIV in REPRIEVE, the PCE risk score underpredicted cardiovascular events in women and Black or African American men in high-income countries and overpredicted cardiovascular events in low-income and middle-income countries. Underprediction in subgroups should be considered when using the PCE risk score to guide statin prescribing for cardiovascular prevention among people with HIV in high-income countries. Additional research is needed to develop risk scores accurate in predicting ASCVD among people with HIV in low-income and middle-income countries. FUNDING US National Institutes of Health, Kowa Pharmaceuticals America, Gilead Sciences, and ViiV Healthcare.