Myocardial Infarction among HIV-infected patients enrolled in the North Carolina Medicaid program

The introduction of combination antiretroviral therapy (cART) in the mid-1990s for the treatment of Human Immunodeficiency Virus (HIV)-infection has substantially reduced AIDS related morbidity and mortality. However, non-AIDS related conditions including myocardial infarction (MI) events are an increasing concern to HIV-infected patients, their providers and the HIV care management system. There has been some evidence from observational studies that suggests other antiretroviral agents, including abacavir, may increase the risk of myocardial infarction, independently of their effect on traditional MI risk factors, however, meta-analyses of clinical trial data has not confirmed these findings. Administrative claims data may be a valuable resource for studying long term and rare outcomes related pharmaceutical treatments and little work on HIV clinical outcomes has been attempted using these types of data in the United States. Therefore, we aimed to further investigate the effect of specific antiretrovirals on MI using the North Carolina Medicaid administrative data. In order to evaluate effects of treatments in the absence of a randomized controlled clinical trial (RCT), it is important to design a study that would most closely an RCT, should it be possible to conduct such a study. Therefore, we first validated myocardial infarction outcomes using the UNC HIV CFAR Clinical Cohort (UCHCC) as a gold standard. We showed that the use of ICD-9 codes combined with length of hospitalization criteria has a high specificity and moderate sensitivity for the ascertainment of MI events (Sensitivity: 0.588-0.824, Specificity: 0.982-0.994). These findings are important as high specificities reduce the potential for bias due to outcome misclassification in comparative safety studies such as this one. We then conducted a new user, active comparator cohort study to investigate the relationship between specific antiretroviral use and myocardial infarction outcomes. We found that the rate of MI among recipients of abacavir with or without zidvoudine as a part of the cART regimen was higher than that of tenofovir (Adjusted Hazard Ratio: 1.43 [95% Confidence Interval: 0.25, 8.31] and Adjusted Hazard Ratio: 2.95 [95% Confidence Interval: 0.89, 9.72] respectively). We did not observe clinically meaningful differences in the effect of other antiretroviral treatments on MI.