Abstract 17992: Observed versus Predicted Myocardial Infarction Among Individuals With Human Immunodeficiency Virus by American College of Cardiology/American Heart Association Pooled Cohort Equation Risk Strata: Center for AIDS Research Network of Integrated Clinical Systems

Abstract

Background: Human immunodeficiency virus (HIV)-infected individuals are at elevated risk for cardiovascular disease (CVD) due to an interplay between traditional CVD risk factors, chronic inflammation persisting despite HIV viral suppression, and side effects of antiretroviral therapy. Previous studies evaluating CVD risk prediction models in HIV populations have generally been small in size and assessed non-U.S. cohorts. The 2013 ACC/AHA ASCVD Risk Estimator has not been evaluated in a large, multi-center HIV cohort. Hypothesis: Our primary hypothesis was that the ACC/AHA ASCVD Risk Estimator would underestimate ASCVD risk across risk strata for HIV-infected patients. Methods: We evaluated this risk prediction tool in the Centers for AIDS Research (CFAR) Network of Integrated Clinical Systems (CNICS) multi-center U.S.-based cohort of approximately 30,000 HIV-infected adults with rigorous central adjudication of myocardial infarction (MI) and differentiation between type I and II MI. We compared MI and ASCVD risks predicted by the ASCVD Risk Estimator to actual rates of MI observed in the CNICS cohort. Results: A total of 132 MIs were observed during follow-up, compared with 103 MIs that would have been expected based on the ASCVD risk estimator. Observed MI rates were higher than expected across most predicted risk levels, particularly among black participants and those with low (<5%) predicted 10-year ASCVD risk. This under-prediction was relatively uniform across risk strata among white men; the observed MI rates were 68%, 68%, 67%, and 67% greater than expected, respectively, for white men with <5%, 5% to <7.5%, 7.5% to <10%, and 10% or greater predicted 10-year ASCVD risk. Conclusions: These data suggest that the ACC/AHA Risk Estimator under-predicts MI risk among HIV-infected individuals. The under-prediction was relatively uniform across risk strata for white men. These data highlight the need for accurate HIV-specific CVD risk prediction tools.