Date of Award


Document Type

Campus Access Dissertation


Epidemiology and Biostatistics



First Advisor

Angela D Liese

Second Advisor

Wayne A Duffus


Objectives: To determine the incidence rate of diabetes, dyslipidemia, and cardiovascular (CVD) events in a retrospective cohort of HIV-infected and non-HIV-infected individuals. And to evaluate the association of these outcomes with exposure to highly active antiretroviral therapy (HAART) medications i.e. protease inhibitors (PIs), non-nucleoside reverse transcriptase inhibitors (NNRTIs), and nucleoside/nucleotide reverse transcriptase inhibitors (NRTIs).

Methods: The study population included a matched cohort of HIV-infected and non-HIV-infected individuals, ≥ 18 years old, who were served through the South Carolina Medicaid program during 1994-2011. The repeated measures of CD4+ T-cell count and HIV viral load were obtained from the statewide SC HIV/AIDS surveillance database. Time-dependent proportional hazards analyses and marginal structural models, which assess the causal association in observational data, were employed in the analyses.

Results: A total of 13,632 eligible individuals (6816 HIV-infected and 6816 matched non-HIV-infected), with a median follow-up of 70 months (IQR: 23-105), contributed 88,359 person-years. The incidence rate per 1000 person-years in HIV-infected individuals as compared to non-HIV-infected control subjects was significantly lower for diabetes (10.89 vs. 13.40; p-value = 0.0007); whereas difference was not significant for dyslipidemia (22.99 vs. 23.23; p-value = 0.818) and CVD events (21.91 vs. 20.16; p-value = 0.070). Time-dependent proportional hazards analysis suggested that after adjusting for confounding factors, HAART treated HIV-infected individuals were at a significantly higher risk of dyslipidemia (aHR 1.18; 95% CI 1.07-1.30) and CVD events (aHR 1.15; 95% CI 1.04-1.27) compared to non-HIV-infected control subjects.

The marginal structural models suggested that among HIV-infected individuals use of PIs versus no use, increased the risk of diabetes, dyslipidemia, and CVD events; whereas the use of NNRTIs and NRTIs versus no use, was causally associated with increasing the risk of dyslipidemia and CVD events. In addition, worsening viro-immunological status and Hepatitis C co-infection were found to significantly increase the risk of CVD events.

Conclusions: The absolute incidence rate of cardiometabolic disorders and CVD events is not currently higher in HIV infected population compared to general population. However, the risk of CVD may substantially increase with disease progression and use of HAART medications, especially the PIs-based regimens.


© 2012, Avnish Tripathi